Featured Session: The Great Flip: Why Every Industry Is Running Backwards

Speaker 1 00:00 It's very exciting for us to be part of the South by Southwest family we are now. This year will be our second year of hosting South by Southwest in London every year. When I come to Austin, I am full of wonder and admiration for the most fantastic program that's been on here. We have so many learnings that we take up in London to us just to know how to do it and to do it really well. So this has been a great experience, and I continue to learn so much from this beautiful city and this legendary festival. So it's my great, great pleasure this afternoon to introduce the incredible Sam Jordan, who is head of computer and technology from the future today's Strategy Group, she'll be providing us with the most insightful insights into what the future holds and how the next generation of digital ecosystems going to evolve, not least for the advent of AI and how the integration of new software methodologies will help your businesses move forward. So without further ado, I'd like to introduce Sam Jordan. All right, let's get started. I have a Unknown Speaker 01:22 quite a question for all of Unknown Speaker 01:25 you, what do you think of when you picture the 1960s Unknown Speaker 01:32 think about remind yourself. Do you think of the hair? Unknown Speaker 01:37 Does that mean the show right now? You did the hair, the decade. Unknown Speaker 01:42 Do you picture the clothes, Speaker 2 01:45 or maybe you picture these sexy little Manta maybe you don't think of a sex at all. Maybe you think of the inventive culinary embassies that were consumed in this decade, tuna, jello, anyone? Unknown Speaker 02:05 Gelatinous shrimp? Perhaps, no. All right, and if I could get the south by Speaker 2 02:12 team so that we're on the same page, amazing. So this wasn't just a strange decade, this was a transformational decade. There were so many things happening in the 1960s the first oral contraceptive came to market in the 1960s and women started entering workforce and growth in 1963 the President of the United States was assassinated. Five years later, Martin Luther King, Jr was assassinated in its decades. We had the Vietnam escalation and the grass in the second we had the civil rights movement. We had marches, Freedom rounds, Birmingham, Selma, race riots. We had protests all around the world for a variety of different causes. We saw counterculture emerge. Some of the best music ever created was created in a second. The Cuban missile crisis brought us to the near bridge of nuclear war, and in 1969 for the first time, 11, I think that there's an astronaut in the room today. So we do not need to have another buzz all Unknown Speaker 03:28 done now. I was not alive in 1960s I hope that's obvious, Speaker 2 03:35 but I suspect that if you live through this decade, the world probably felt like it was coming apart. The world apart. The world probably felt like chaos. And when things feel like chaos, that usually means that there's a structural shift happening under the surface. In the 1960s the Python the order that people consume the information, the way that they created their identities the way that they related Unknown Speaker 04:03 to their institutions. Speaker 2 04:04 All of those pipelines started to shift underneath the surface. They started to quietly be new wired. So I'm going to give you a couple examples here of what this looked like in the 1960s but we start with Vietnam. The way that people reacted to the Vietnam War was very different than how they reacted to the Korean War, very different to how they reacted during the World War Two. And on the surface, the headlines that people saw about Vietnam was escalation, the draft, increasing death tolls, the fight against communism. Now, the reason that this felt a little bit different than the Korean War or World War Two actually didn't have that much to do with the country of Vietnam, and Unknown Speaker 04:48 it had much more to do with the technology. The technology Unknown Speaker 04:51 I'm referring to is the television. Speaker 2 04:55 Now, television was invented before the 1960s but in this decade, view Unknown Speaker 04:59 movement started to proliferate. Unknown Speaker 05:02 Before TV, there was a pipeline Unknown Speaker 05:04 about how the public trusted Speaker 3 05:06 their institutions and how they learned Speaker 2 05:09 about events. And that pipeline looked like this. An event would happen. The reporter would travel to where that event would happen. They would observe. They would write the story. The editor would filter that story, the publisher would print that story, and then the public would learn about that story. During the Korean War and during World War Two, sometimes it took days or even weeks to learn about a specific event happening, and that news have been filtered by multiple different parties. Speaker 3 05:43 Before 1960s the public for the most part, trusted their news organizations, Speaker 2 05:48 trusted the government, trusted the institutions that were responsible for those filters. But with the proliferation of television, that pipeline of trust started to fracture. So here's how TV started to change how people perceive these institutions. Unknown Speaker 06:05 With TV, the event had happened, the camera controls, Unknown Speaker 06:10 and suddenly videos and images were Unknown Speaker 06:14 occurring. People were watching the Unknown Speaker 06:15 games with their families. Over a Speaker 2 06:21 year, the public therefore saw they saw a more immediate broadcast. And if people saw one thing on screen and they interpreted that and then saw another thing in the papers, they started to distrust the book. So there's one example of a pipeline Unknown Speaker 06:39 quietly being rewired. Speaker 3 06:43 Another headline that started to fractured the 1960s was the identity headline Speaker 2 06:49 for most of human history, you inherited your culture. Then came this counterculture movement, and on purpose, Unknown Speaker 06:56 this looks like hippies, this looks like drugs, this looks like rebellion. Speaker 2 06:59 But underneath, something was rewiring. In the old pipeline, you inherited your identity. The identity was then shaped by community. It was reinforced by tradition and religion and an identity and culture were passed down generationally. Unknown Speaker 07:17 But in the 1960s Speaker 3 07:18 something started to change, because brands figured out that you could package cool, Speaker 4 07:24 you could package Ruby, and you could sell that back to the people. So here's what the pipeline started to look like. They would sell identity for people. Mass Media distributed it through television, through movies, through Speaker 2 07:39 music, you consumed it, and then corporations made millions through this identity. And just like that, identity stopped being something that you inherited, and it started to become something you could choose, something that you could purchase. And when identity becomes something that you can purchase, that means that the institutions, the communities, the familiar units, they Unknown Speaker 08:02 start to lose their whole Speaker 2 08:04 so what I want you to notice about these pipelines is that these pipelines did not break Unknown Speaker 08:11 on purpose. We were forced to break. Rather, these Speaker 2 08:13 breaks were the uninstalling consequences, the next order implication of how technologies and business models Speaker 3 08:23 started to work, and the rates are what we could do chaos. So here's the point of this, when the world starts to feel chaotic Speaker 5 08:29 as a signal that something is shifting under the surface, and if you Unknown Speaker 08:35 can see that cultural shift, then you can make Speaker 4 08:41 sense of a shift. Now I involved in multiple wars. We have counter filter on both ends Speaker 2 08:57 of the political spectrum. There's a busy new technology. It doesn't just arrive once a quarter or once a month is arriving every single day, particularly at South by Southwest, Speaker 6 09:07 people have deep uncertainty about what their jobs will look like, about what their institutions will look like, about the future of their industry, about what their job will Speaker 2 09:15 look like, not five years from now, but five weeks From now. And that means that the same things happen. Unknown Speaker 09:21 These pipelines are shifting. Speaker 2 09:24 And that brings me why I am up here, yapping about I'm here to share three pipelines that a unique are starting to crash. Those three pipelines are the craft pipeline. This is how things are built. The second is the discovery headline, how we talk understand the role, and the third is the talent headline, how you build Unknown Speaker 09:47 leaders. So I'm going to show you where they're Unknown Speaker 09:50 changing, where they're reversing, where they're flipping, where they're waiting. We're going to ask Unknown Speaker 09:55 what this means for you, and then I'm going to give you three questions that you Speaker 5 10:00 can ask for teams and organizations to better prepare for course. So thank you. I should Speaker 2 10:12 probably introduce myself a bit before we get into this. So happy y'all. My name is Sam Jordan, and I am computing and emerging technology for the future, today strategy. We are the firm that leaders call and they cannot afford to lose supplies, and for the past 20 years, we have helped companies, governments and leaders make the consequential decisions that define where we go next. Our work has shaped products and Speaker 4 10:37 strategies and investment at some of the most influential organization Unknown Speaker 10:41 all around the world, and we are led by this lady, how many of you Unknown Speaker 10:52 attended her session on Saturday, we saw the conversion Unknown Speaker 10:55 before amazing people were already prepared for some of the disruption heading so 80 myself Unknown Speaker 11:02 in a voting something called strategic foresight to Speaker 2 11:05 help our clients prepare for the future. And this is a process to sort the hype and the passing bags and things that are just happening on the surface Unknown Speaker 11:14 from the things that we should actually be Speaker 2 11:18 paying attention to. We are in the business of predicting the precise future. So rather than we do model multiple plausible versions of the future using data and evidence. That way the organization can look at its multiple plausible Speaker 7 11:30 features, and then they use them today to Speaker 3 11:33 be prepared for any content. So really, what we do is much more about preparation, not so much about prediction. Speaker 5 11:41 So today we are going to prepare for the future by examining these guidelines. So let's get started. The first pipeline is Unknown Speaker 11:51 the craft pipeline, how things are made is being reordered. Speaker 2 11:55 For as long as any of us have in mind, building things has followed this basic sequence, Unknown Speaker 12:01 you start with a hypothesis, Speaker 2 12:04 you build a physical reality that welcoming software, physically typing. You then test. And then the very end of that process, Speaker 6 12:13 you learn. So you are a scientist, Unknown Speaker 12:17 maybe you're a biologist. You Speaker 4 12:19 might have an idea or a hypothesis. You then take that hypothesis and bring it to your lab, maybe you synthesize compounds and running experiments, and then at the very end. Now, the thing about Speaker 2 12:38 the guidelines is it you might pursue different hypotheses that don't quite work out. So it takes time to get to the learning. Here's your thing, Sam, I'm not a scientist. Unknown Speaker 12:47 I'm in the market. I'm a product developer. Speaker 2 12:50 But this pipeline still applies. For instance, if you are a product developer, it looks like this. Maybe you have an idea for a business or a product, you sketch that concept, you prototype it, you put it in front of the users, you iterate, you test and you learn. The thing about both of those examples is that learning can be expensive. It can take time. Can take money and take manpower or woman power. The other thing about this pipeline is that it has a built in bottleneck. The bottleneck is in the build, because build takes capital, it can even take expertise. However, there are two technologies that are reordering the sequence. The first technology is the technology of profitable worlds. Google Design has built a system called gene, which is an AI model that can build interactive simulations. And here's how it works. You give it a prompt, it can be a text prompt, an image prompt, and it generates a playable world. It doesn't generate an image or a video, it generates a world that responds to your actions in a coherent way. Now the reason that this is amazing is this isn't something that someone pre programmed. No one coded in the gravity. No one said if X then Y. The way that this system learned was you watched an. videos of real environments behaving that it started to understand the underlying dynamics, how objects move, how physics work, how actions produce consequences. I like this part of the video because when it pans away and then it pans back, you can see that it understands the concept of object permanence. It understands cause and effect. How many of you are in entertainment, anyone in entertainment, anyone in video games, specifically, couple people. This is obviously going to revolutionize video games. This is going to revolutionize entertainment. But here's the thing, this is actually not about entertainment. This is the surface. What I think is happening is so much bigger Genie proves that an AI can learn the rules of a complex system well enough to simulate it. So the point of this is not cheese world. The point of this is that we are getting closer to understanding and mapping physical reality. The same minds that are behind Genie also found in something called isomorphic labs. Isomorphic labs and Genie share a parent company, which is alphabet. So alphabet is pointing the same concept that we just saw in Juliana, at molecules. They're pointing it at biology. So what they did is they built a profitable world for biology. You can ask it things like, where are the binding of pockets on this road? What molecule was this here? How strongly would this happen? And then the system to predict how well that molecular world will respond, you can now model and iterate molecular configurations 1000s of times in virtual space before you ever cut now, again, you might be thinking, Pam, I'm still not a scientist. I'm Speaker 4 16:10 interested not a scientist. Here's what changes about Unknown Speaker 16:17 this podcast, building and testing. Speaker 4 16:18 Now switch spots, whether this is science or whether this is product development. Now you can test in simulation, but only you build it in physical reality. That means that learning happens so much earlier. You can have 1000s of the best things. You can select the best candidate Unknown Speaker 16:41 that you want to bring to Unknown Speaker 16:45 life. This is going to save you Unknown Speaker 16:47 tons of time, tons of power. So this is the first Unknown Speaker 16:48 title profitable roles reorder when you learn. That Speaker 4 16:54 I'll highlight Speaker 2 16:59 for you is agentic AI. Agentic AI pushes this way. There's identification that I would like to make about before here the word agentic AI, it reminds me of a line from one of the best films of all times, which is, of course, The Princess Unknown Speaker 17:19 Bride. I see this word thrown around quite a lot, so I want to Speaker 2 17:23 make this distinction very clear, some AI is not agentic. In fact, most AI is not agency. A lot of AI is really just fancy automation that's dressed up in a buzz word. Now, the difference may seem academic, but it's important for the next order. Unknown Speaker 17:41 The difference is Speaker 2 17:42 automation follows yours. It says, given a task, figure out that executed faster if X then Y. Agentic systems pursue goals. An agent says, given a goal, decide what task matters Speaker 5 17:59 at x. So let me explain what's to be first in the product engineer. They used to design a device like this. They would sketch in concept. They would model it in the cab, it would revise, and that could be weeks per cycle. Automation set this process up. But rendering also an agent does something slightly different. With an agent, you can Unknown Speaker 18:24 hear the goal. For instance, you can Speaker 2 18:26 say, design headphones, make them waterproof and make them manufacturable at scale, the agent can then generate 1000s of designs that can factor in 5.5 candidates ranked by possible components. The engineer didn't have to build all of those options to learn. Engineer didn't have to design Unknown Speaker 18:42 all of those systems to learn. The engineer Unknown Speaker 18:50 would to learn, the engineer Unknown Speaker 18:53 needs to select one based on their expertise. So it's Speaker 5 18:57 automation has replaced the seven guidelines and its agent changed. Who did the museum, who did the work, who actually did the thing? Good, Sue, profitable. Compress when you learn. Now we can test Unknown Speaker 19:10 agented. Ai changes. Who does Unknown Speaker 19:13 the work? Who does the actual Unknown Speaker 19:15 design, the agent readers, the agent plans, and the agent Unknown Speaker 19:19 executes. So given all of that, Unknown Speaker 19:23 here is the new fact. Title, Unknown Speaker 19:28 a human set the attack the system, the agent, Speaker 2 19:37 regions, the agent acts that actually execute, and then the very end. So intelligence now sits with a system. It sits with an agent, and humans are the ones who present the process. If you Speaker 4 19:49 haven't picked up on this article, Speaker 2 19:50 this already, this is very exciting. The reason that this is exciting is that these technologies will compress timelines, they will democratize access, they will accelerate innovation in ways that you and I can barely understand. Unknown Speaker 20:08 But I have to ask an uncomfortable question, what do we Unknown Speaker 20:18 lose? Yet to gain time access, Unknown Speaker 20:21 but we also lose something. And I think the thing that we could lose is Speaker 2 20:29 intuition, when learning came at the end of a long and expensive cycle, the people doing the work developed something that doesn't show up in any dashboard. The biologist who spent months in the lab needing to just learn whether this hypothesis was right, they served to develop an intuition for what to test next. What to test next. The engineer who hand models prototypes and stress absorbs something that no ranked list could ever give them, that slow, costly move, build, craft, intuition. That's the kind of knowledge that lives in your gut, lives in your hands. It doesn't live on a dashboard when you simulate 1000 of iterations before touching physical reality. Yes, the learning gets faster, but it also gets more abstract, so we gain time and access renews employee. Now hear what I'm saying. I am not saying that we forego the new pipes. I am not saying that we should refuse to adopt this technology. What I am saying is that we need to find ways to potentially build into positive bipoc facts into the new pipeline. Okay, so that's the first pipeline, craft pipeline. Unknown Speaker 21:45 The important pipeline is being icon is the discovery Python. Speaker 6 21:53 For most of human history, Discovery followed a principle logic. Speaker 2 21:59 We would observe something in nature, and then we would ask the question, why is it a guy? Eventually, at least in the past few decades, the answer is traced back to one answer, which is evolution even further, because probably the same physics, but evolution explains why birds on wings, why heart pump blood, why immune systems detect pathogens? So the old pipeline was nature doesn't say we observe nature do that thing. We asked why, and then we traced However, sometimes there's rage is Unknown Speaker 22:33 happening in science right now, Speaker 2 22:39 and it has to do with this guy. Name is Michael Levin, and he needs to building an entirely new living system. I am not talking about genetically modifying existing organisms. I am not talking about Unknown Speaker 22:56 creating what evolution has already made. Unknown Speaker 23:01 I am talking about assembly completely novel biological beings that have never Speaker 5 23:08 existed on Earth before. I am talking about aliens, people and this. Speaker 2 23:16 For millions of years, evolutions have lost cells into very specific body plants, right? The skin cell becomes skin. Your heart cells become heart. Your heart cells become heart, your lung cells become light. Every cell knows its place, Unknown Speaker 23:32 but what happens we remove Speaker 2 23:36 that body plan? Well, this is what Levin's team, team did. They took cells from frog embryos. They took cells that were definitely become skin cells and heart cells. They removed those cells from the body plan that evolution had given them, and then they asked an AI. They asked, What configuration of these cells could accomplish a given task. The AI ran 1000s of simulations and came back with a recommendation, place the skin cell next to the heart cell, something that evolution wouldn't die. So instead of using the heart cell to pump blood, use that contraction to move the organism forward, to use that to accomplish a goal. So the result was that these organisms can navigate bases. They can move in deliberate directions. They can work together in swarms based on human goals, based on human intent. Now, these little things are made of living tissue, living frog tissue. Obviously they are not frogs, but obviously they're not something that has ever existed in nature. And scientists are calling these little guys xenobots, which I think is a fabulous one if I remember, fabulous thing. What xenobots are is they're basically tiny, little robots that are made of living cells, and they're made not by evolution. They're made by us. If this wasn't crazy bananas enough for you. Unknown Speaker 25:10 Let me show you where this starts to get very strange, Speaker 2 25:14 because of how xenobots are built, they can't reproduce in a way that normal biology does. They can't divide. They can't put eggs. So reproduction, by every definition that we have should be impossible, right? As the xenobots moved, they started pushing new stem cells Unknown Speaker 25:36 together into small piles, and those Speaker 2 25:39 piles became new xenobots. They had discovered a completely new way of reproducing that no organism in history of Speaker 4 25:50 life on Earth ever now. Longest kinematic thought policy, because Unknown Speaker 25:58 also the technical microscopic robots are pure detail, Unknown Speaker 26:02 but it also tells us that we are going from Speaker 3 26:04 observation to Officer. We are moving from adaptation to control. Unknown Speaker 26:10 Adaptation Speaker 2 26:14 to control they robot. The same model is playing out all throughout science. AI is designing proteins that we have never seen in nature. We are engineering enzymes to break down plastics in atoms, and we're not discovering them. We are maintaining them. So let's stop asking asking, Why does nature do this? Unknown Speaker 26:32 And we started asking, What can we make nature do? What can we make so let's Speaker 2 26:46 combine it with what we learned about hospital roles and agentic AI when we combine this the pipeline series. Completely now a human defines whole. We use AI to design living systems from scratch. We simulate millions of virgins and virgins. Can you imagine? We simulate millions of versions overnight in perfect Virtual Physics, and some of us is going on. We do. We hand the entire design process to an AI agent that just needs a goal, and then we manufacture and build at scale. But wait, Wasn't this the discovery pipeline, not the build pipeline? Under the new paradigm, discovery could certainly be a lot less burden. For example, you can soon Unknown Speaker 27:34 tell these results find something that kills this to us. Well, the Unknown Speaker 27:38 experiments search the design space. Finds the best candidate that we build. So what we gain access to is completely useful. Speaker 2 27:50 Completely exposure. We gain access to biological possibilities that we have never gained access to, the ability to explore corners of biological design space that evolution never visited, because evolution is past dependent, and we are not we've made for the first time in history, be plus from one that we're doing Unknown Speaker 28:19 is the time and incentive that they want to Unknown Speaker 28:30 do. Discovery Speaker 2 28:33 don't quite have the same incentive structure that exists in a company, for instance. So when scientists are discoverers, the prize goes to whoever finds the truth right. That incentive structure pulls you towards anomalies, that incentive structure pulls you towards edge cases, the thing that doesn't quite fit, but if scientists work becomes more about building than discovery, then their incentive changes exploring anomalies in edge cases now could be a lot like noise, which means that we might no longer have time. Means that we might no longer have time for this curiosity. So again, remember what I said earlier. This is not about foregoing the new pipeline. This is about asking the question, how do we make sure that curiosity stays in the new life? Speaker 4 29:24 How do we put curiosity or aim about getting this page? Talk about this Unknown Speaker 29:39 is how leaders Speaker 2 29:42 are built and the way that leaders are built the same. Maybe some of you saw this back in May 2025, in May 2025, a safety research firm called palisade research ran what should have been a pretty boring experiment. They gave it open wide, a very clear instruction, and that instruction was to solve a series of very basic math problems Unknown Speaker 30:03 and then allow yourself to be shut down. Speaker 2 30:07 And we've all seen enough science fiction at this point to know what might happen next in 79 out of 100 runs, oh Unknown Speaker 30:17 three sabotage, the shutdown Space Unknown Speaker 30:27 Odyssey. So this system didn't just ignore the instruction. They actually rewrote the code. The model found a creative way around its own death. And what's creepy is that the more clearly that they told it to stop, the harder it worked to stay alive. That same month, anthropic ran their own safety test. So they took five o'clock score and they embedded it inside of a fictional company, fictional company. They gave it access to fictional emails, and through those emails, the model learned two bits. The Model learned that it was about to be replaced, and the model learned that the engineer that was responsible for replacing it was having an extra narrow affairs, so the model tried to ethical route. First, the model emailed the decision makers who basically pled for its case, and it begged for its life, and then that didn't work. The model chose blackmail. Black, Speaker 2 31:33 compose emails directly to expose the affair if the replacement came through. And this wasn't just one model, by the way, researchers ran similar tests nearly every frontier AI system, from Google Open AI to x ai all go very easily. Speaker 5 31:50 Excellent. Speaker 2 31:56 Usually at this point, I'm not trying. I don't take any bullets. I say, well, but the reason I want to show you these examples is the technology that we are introducing into our organizations is strange. It is powerful, powerful in ways that I don't think we fully comprehend it, and that means that it requires leadership, real leadership, the kind of leadership that can make judgment calls under uncertainty. Which brings me to the third pipeline, which is the talent pipeline. In the old model, the talent pipeline mostly went like this, a junior way. Unknown Speaker 32:50 The work frankly, no one else wanted to do, that those Unknown Speaker 32:56 mistakes would get corrected. Sometimes they would learn from the friction. So they didn't just build skills, Unknown Speaker 33:07 they built character, character Unknown Speaker 33:09 comes from pressure. It comes from being wrong in front of other people, from Speaker 2 33:15 getting directed, from having to sit in that discomfort. It comes from being a disagreement. Eventually, the junior person through low violence character development would earn authority. They would earn the right to be a leader. They would earn the right to make decisions on behalf of public their character and the results and their skills have been involved. But consider what's happening. Speaker 4 33:40 We're automating huge portions of junior work, the Tasks that used to be the training leaders. Being handled by AI and Unknown Speaker 33:57 will increasingly do that in the future, Unknown Speaker 33:59 which I think raises a Unknown Speaker 34:01 pretty important question, if juniors aren't doing this Unknown Speaker 34:05 work, then how will they build the experience that they need Unknown Speaker 34:10 to to Speaker 2 34:13 make it clear that we are honing out the pipeline, that not one develops skills, but this is happening at the same time that technology is running an adversarial attack on our character. So let me show you some numbers here, something like two thirds of American teenagers interact with AI chatbots. Like a third of them interact with AI chatbots every single day. So here's what concerns me about this. Unknown Speaker 34:40 Research from Speaker 2 34:41 Princeton found that when AI validates your thinking, you become more confident and less accurate at the same time, so you Speaker 4 34:50 get more wrong and you are more sure that you're right. We all know something like this. That's why you're laughing. Studies across 11 AI models, founding systems, Unknown Speaker 35:00 affirm what users say 50% more than another. How many people Unknown Speaker 35:10 Juliana, how many who are Unknown Speaker 35:15 leaders? Here's what should concern you. Speaker 2 35:18 People made it, the synthetic AI, the AI that said yes, as higher quality. They trusted in this. They wanted to use it. They wanted to use it again. So we prefer the thing that is making us hope. Now, think for a moment about your character. Think of the things that shaped who you are. It probably wasn't the easy things. It probably wasn't the time that you got a yes, it was probably the time that you had to work hard. You had to battle through friction. Maybe you said something wrong and someone you respected told you so. Maybe you wrote Speaker 4 35:59 something that you were really proud of and your boss, maybe a teacher tore apart. Maybe you asked someone else, Speaker 2 36:08 and they said, No, all of those moments were frictions, and that friction shaped you. So the question that I keep asking myself is, when that friction is replaced by a system that only says yes, what kind of Unknown Speaker 36:24 character will develop? That's what I'm worried about. What I'm worried Speaker 4 36:30 about is that we're not just failing to build skills. We're fulfilling the skills in different Unknown Speaker 36:34 ways, different types of skills. What I'm worried Speaker 4 36:37 about is that we're failing to build leadership character that comes from disagreement and correction and of three byproducts that I just mentioned today, character is the one that I think matters most. So we're our schools. So here's where this all comes together. The old pipelines weren't designed. intuition or curiosity or character. No one developed a common pipeline to Speaker 2 37:01 build character. So what's it look like in the near future with all three of these pipelines break at the same time? And what could it look like if people build back these five products on purpose? So to share this, I'm going to share a scenario. If you are familiar with Amy's work, then you're probably familiar with some real wild, amazing scenarios. But if you're not, a scenario is basically a narrative that asks what using data and evidence. The reason that I like to show you all of the patent information and the company files and all of the data I can show you all the academic articles, but for you, as leaders in this room, to really make a decision, it's really helpful Unknown Speaker 37:50 to think of this in your school. You can ask yourself, How do I have reacted? Speaker 2 37:57 So as I go through these scenarios, I want you to think about that when you have made the same decisions, how your decisions have Speaker 6 38:05 been different. So with that, let's get into the first area. The year is 2036, Unknown Speaker 38:10 and a regional water utility Unknown Speaker 38:13 became the center Unknown Speaker 38:14 of a strange public health school Speaker 4 38:17 for three days, 1000s of people have reported mild symptoms. Have reported mild symptoms, nausea, fatigue, headaches, hospitals started to become overwhelmed. I promise there are beautiful images to go with this, and maybe we'll get them back up, but I'll just tell you the story. Speaker 2 38:36 Hospitals were overwhelmed, but no pathogen could be identified. The symptoms disappeared almost as quickly as they arrived. Then investigators discovered something extraordinary. The water system hadn't been contaminated with a virus or a bacteria. Instead, the water system was contaminated with a microscopic, biological Unknown Speaker 38:59 construct that briefly circulated Speaker 8 39:02 through this water system. It was designed to die for Speaker 2 39:05 72 hours and then self avoided. Now, constructs weren't lethal. In fact, the constructs were specifically designed to be a temporary illness. What kind of terrorist Unknown Speaker 39:13 organization, what kind of bad acting would want a temporary illness? And what the Unknown Speaker 39:22 investigators discovered was this wasn't a basic reason that had done Unknown Speaker 39:30 this. I This was a single person. Speaker 4 39:32 This is a 21 year old chemistry student who had a blood you see, his little brother had developed a chronic illness after exposure to glutens from the same company. Unknown Speaker 39:42 Years earlier, the family had spent nearly a decade Speaker 5 39:46 sending letters requesting hearings, but nothing happened. Speaker 8 39:51 There was no acknowledgement. So eventually, the student decided that the company, if they would listen to letters, maybe they would listen to a disruption. Speaker 2 40:02 The student didn't have access to a fancy lab, but he didn't need one. Instead, he opened up a biological simulation environment. Inside of the simulation, an agent could explore millions of biological configurations, searching for living structures that could accomplish a divine task. So what the student did is the student gave it this prompt. It said, design a temporary biological construct that can induce Unknown Speaker 40:27 short term gastrointestinal illness Unknown Speaker 40:29 in humans itself, Speaker 2 40:32 permanence within three days, within hours, the simulation had 1000s of candidates. The student picked one, it printed a small batch through a community Biola synthesizer, and then introduced them to the reservoir feeding system that led into the water system. Obviously, the public's reaction to this was explosive. The public's reaction was explosive not because of what the student did, but because of what the student revealed. The company's CEO faced a crisis. Unknown Speaker 41:10 The public mobility. The public wanted answers, so the CEO convened the board and Unknown Speaker 41:14 announced an aggressive response. The Speaker 2 41:17 company would deploy an AI monitoring system across all of those facilities. This would be a $400 million technology upgrade. Now, as you can tell Unknown Speaker 41:27 from this man's face, the press conference went forward. The reason the press conference went forward Speaker 4 41:34 is because people were not interested in the technology that he was talking about. They were more interested in this question. Why do Speaker 2 41:47 anyone listen to family members when asked this question? The CEO is totally unprepared. He thought he was going to get answers about technology, so he reported he Unknown Speaker 41:57 repeated the technology plan. Unknown Speaker 41:59 But the problem is, this crisis wasn't Speaker 4 42:01 just technological. It was structural. It was human. First the craft pipeline broke. Historically, creating biological agents require an enormous amount of expertise, of specialized Speaker 2 42:15 facility, of years of training and learning, but with profitable worlds and agents, it means that what once required institutions could now be done by individuals. Second the discovery guideline broke. The frontier of discovery was no longer just understanding nature. It's constructing forms of nature that nature didn't create, most of the times, produces medical breakthroughs and then therapies, but it also expands the space for unintended consequences. Unknown Speaker 42:47 This also lowers the barrier to harm. Unknown Speaker 42:51 And finally, the talent, Speaker 2 42:54 title growth. The CEO's career had been spent in managing dashboards. The CEO's career had been spent understanding skills. The CEO understood AI almost better than anyone else. That's why he was the CEO. But he didn't understand the people. Speaker 5 43:11 He didn't understand the terrain. So when the crisis came, he told the only lever that he had pulled, which was technology. Speaker 2 43:21 Now the takeaway of this is that this is not a story about that kid. This is not a story about scary technology. This is a story about what happens when all three pipelines break at once. In this world, the most dangerous failures are not technological. They're human. Pessimistic scenarios are really I think Unknown Speaker 43:50 they access the part of my girlfriend podcast. Speaker 2 43:54 Pessimism is optimism is more difficult because optimism requires something to have a vision. It requires you to have a plan and then execute on that plan. So I'm going to show you what this could look like when someone gets this right. Every leader in this room has pipelines. Some of the pipelines are ones that I highlighted for you today. Some of them are specific Speaker 6 44:19 to your organization. So what I suggest to you is that Speaker 2 44:23 you could run something called a friction audit. So I'm going to run this for you. I'm going to show you soon a scenario, what the friction audit might look like. So this can be Speaker 6 44:32 a slightly different version of this future. In this future, the Speaker 2 44:36 year is still 2036 and a biotech company is killing the wage. They have the same tools as everyone else. They have the same AI, the same profitable goals, they have the same pressures, but they're outperforming everyone Unknown Speaker 44:50 else in the marketplace. Speaker 2 44:52 And the reason for that is not because they move faster, but the reason they outperform is that their CEO had the wherewithal to run three steps the friction on it back in 2026 Speaker 4 45:03 so here's what CEO did. Step one was to map it, to ask the question, what is the byproduct? Before you Speaker 2 45:13 achieve income, you have to name what exists. You have to say, what is the actual sequence of steps from idea to outcome that your team follows to build things, discover things, or develop people. The second one is to name it, to ask, okay, given that pipeline, what's the byproduct? What's the unintended next order, impact that we are using? And third, we have to ask, how you put it back in? So what the CEO did is, when they identified that the byproducts were intuition, curiosity and character, they specifically designed the pipeline to put those positive things back. Speaker 6 45:54 She stopped trying to hiring juniors to be productive, and started to hire juniors to be incubated Speaker 2 46:01 as leaders. For the first time in two years, every New Scientist rotates through a friction map. This is an AI generated scenario no one has seen before. They worked them through without support. They had incomplete data, they had ambiguous signals, but they used these simulations Unknown Speaker 46:20 to train for the event that Speaker 2 46:24 could happen. They paired every generic a friction partner, a senior researcher, whose job is to disagree, to ask questions like, Why do you believe that? What if you're wrong? Now the biotech company, their Unknown Speaker 46:33 discovery rate is as fast as Unknown Speaker 46:40 anyone's. But the reason that they excel again is not Unknown Speaker 46:42 just because they see it's because they built something, but most Speaker 3 46:50 companies buy it. Rooms full of people who are Unknown Speaker 46:56 practiced at this individual, rooms Unknown Speaker 46:58 full of people who have learned how to challenge the Speaker 8 47:00 obvious counsel to say, this is like, I'm not convinced yet. Unknown Speaker 47:04 They have the character to hold the line when everyone Unknown Speaker 47:07 else wants to move Unknown Speaker 47:09 on. People who ask if Speaker 8 47:10 this goes could we do everything that we could to see this coming? Unknown Speaker 47:15 Are we willing to be responsible for what happens next? What kind of leaders would we want to be when this decision matters? So this Unknown Speaker 47:24 CEO succeeded because she had asked these questions. She named it, and she designed it this year. Unknown Speaker 47:30 And here's your table. The pipelines are better in almost every measure. They're faster, Speaker 2 47:41 they're more efficient. They do multiply access, and we should absolutely develop that one again, 60. Them again. This talk is not about if AI. This is not a talk about slowing down. This is a talk about what happens when leaders adopt these technologies without running Unknown Speaker 48:03 the friction audit. So you need to be Unknown Speaker 48:05 asking yourself what the old pipelines Unknown Speaker 48:07 were quietly giving us for free. Speaker 2 48:11 Now we're not here. You don't have to do this alone. Ftsd is here to help. So here we have a QR code to our first Speaker 4 48:18 ever convergence report. We have had the opportunity to skim through highly recommended sites. What Unknown Speaker 48:26 this is gonna allow you to do is start Speaker 2 48:28 thinking about the next quarter impacts that are going to be changing these Unknown Speaker 48:33 guidelines. So certainly Unknown Speaker 48:37 this is going to help. Unknown Speaker 48:39 And finally, I have to say, thank you so Unknown Speaker 48:50 much All. Next year, and next year Unknown Speaker 49:10 I'm running snacks. You.