# Transcript: Featured Session: The Great Flip: Why Every Industry Is Running Backwards **Date:** March 16, 2026 · 10:30 PM **Session:** [Featured Session: The Great Flip: Why Every Industry Is Running Backwards](/sessions/2026-03-16/pp1148588-featured-session-the-great-flip-why-every-industry-is-running-backwards) ## Summary Sam Jordan from Future Today Strategy explores how three fundamental pipelines are reversing in the AI era: the craft pipeline (how things are built), the discovery pipeline (how we understand the world), and the talent pipeline (how leaders develop). Through examples like AI-simulated worlds, xenobots, and self-preserving AI systems, she demonstrates that while these technologies bring massive gains in speed and access, they risk eroding critical byproducts like intuition, curiosity, and character that leaders need. ## Topics `artificial intelligence` · `digital transformation` · `leadership development` · `strategic foresight` · `agentic ai` · `biological simulation` · `organizational change` · `talent development` · `innovation pipelines` · `business transformation` ## Key Takeaways 1. The craft pipeline is flipping: AI-powered simulations (like Google's Genie) and agentic systems now allow testing before building, compressing learning timelines but potentially eroding hands-on intuition that comes from slow, costly work. 2. The discovery pipeline is reversing: Scientists are moving from asking 'why does nature do this?' to 'what can we make nature do?', using AI to design novel biological systems like xenobots that have never existed in evolution. 3. The talent pipeline is hollowing out: As AI automates junior work and validates thinking without friction, organizations risk losing the character-building experiences (disagreement, correction, failure) that develop true leaders. 4. Leaders must run 'friction audits' to identify what the old pipelines gave us for free (intuition, curiosity, character) and intentionally design those byproducts back into new AI-powered processes. 5. Organizations that succeed won't just move faster with AI—they'll build teams practiced at challenging assumptions, holding the line under pressure, and taking responsibility when decisions matter most. ## Full Transcript What do you think of when you picture the 1960s? The hair? The clothes? The sexy little sedans? Maybe the inventive culinary delicacies like tuna jello or gelatinous shrimp? This wasn't just a strange decade—it was transformational. The first oral contraceptive came to market, women entered the workforce in growing numbers, President Kennedy was assassinated, followed five years later by Martin Luther King Jr. We had Vietnam escalation, the civil rights movement, marches, race riots, protests worldwide, counterculture, the Cuban missile crisis, and the moon landing. If you lived through the 1960s, the world probably felt like chaos. And when things feel like chaos, that usually means there's a structural shift happening under the surface. The pipelines—the order that people consumed information, created identities, and related to institutions—started to quietly rewire. Vietnam provides one example. People reacted to Vietnam very differently than to the Korean War or World War Two. It wasn't really about Vietnam itself—it was about television. Before TV, news followed a pipeline: event happens, reporter travels, observes, writes, editor filters, publisher prints, public learns. This could take days or weeks. People trusted their institutions. With TV proliferation in the 1960s, that pipeline fractured. The camera captured events, and suddenly videos and images were broadcast directly into homes. People watched with their families. When they saw one thing on screen and read something different in newspapers, they started to distrust institutions. The pipeline of trust was being rewired. Another pipeline that fractured was identity. For most of human history, you inherited your culture from community, tradition, religion—passed down generationally. But in the 1960s, brands figured out they could package cool and rebellion and sell identity back to people. Identity stopped being inherited and became something you could purchase. When identity becomes purchasable, institutions and communities lose their hold. These pipelines didn't break on purpose—they broke as unintended consequences of how technologies and business models worked. The chaos was a signal that something was shifting under the surface. Today, we're experiencing similar chaos: multiple wars, counterculture on both political ends, constant new technology arriving, deep uncertainty about jobs and institutions. The same thing is happening—pipelines are shifting. I'm Sam Jordan, head of computing and emerging technology at Future Today Strategy. We're the firm leaders call when they cannot afford to lose. For 20 years we've helped companies, governments and leaders make consequential decisions. We use strategic foresight to sort hype from signals that matter, modeling multiple plausible futures so organizations can prepare rather than just predict. Today we'll examine three pipelines being reordered: the craft pipeline (how things are built), the discovery pipeline (how we understand the world), and the talent pipeline (how leaders are built). For as long as anyone remembers, building things followed this sequence: hypothesis → build (physical reality, software, typing) → test → learn. Whether you're a scientist synthesizing compounds or a product developer prototyping, learning comes at the end of a long, expensive cycle. Two technologies are reordering this sequence. First, predictable worlds. Google DeepMind built Genie, an AI model that generates interactive simulations from text or image prompts. It creates playable worlds that respond to your actions coherently. No one programmed the gravity or coded if-then statements. It learned by watching hours of videos of real environments, understanding underlying dynamics—how objects move, physics work, actions produce consequences. Genie proves AI can learn rules of complex systems well enough to simulate them. The point isn't game worlds—it's that we're getting closer to mapping physical reality. The same minds founded Isomorphic Labs, pointing the same concept at molecules and biology. They built a predictable world for biology where you can ask about binding pockets, molecular interactions, and the system predicts how molecular configurations will behave. You can iterate thousands of times in virtual space before building anything physical. This changes the pipeline. Now: hypothesis → test in simulation → build in physical reality → learn. Building and testing switch spots. Learning happens much earlier. You can run thousands of iterations, select the best candidate, then bring it to life. This saves enormous time and resources. Predictable worlds reorder when you learn. The second technology is agentic AI. I need to make a distinction: most AI is just fancy automation dressed in buzzwords. Automation follows instructions—given a task, execute it faster. Agentic systems pursue goals—given a goal, decide what tasks matter. An example: product engineers used to sketch concepts, model in CAD, revise over weeks per cycle. Automation sped this up. An agent does something different. You give the agent a goal: design waterproof headphones manufacturable at scale. The agent generates thousands of designs, factors in physics, ranks candidates by performance and cost. The engineer didn't build all those options to learn. The engineer selects based on expertise. Automation replaced steps. Agentic AI changes who does the work—the agent researches, plans, and executes. Here's the new pipeline: human sets goals → system (agent) researches options → agent designs and executes → human approves. Intelligence now sits with the system. Humans curate the process. This is exciting—these technologies compress timelines, democratize access, accelerate innovation. But we must ask an uncomfortable question: what do we lose? We gain time and access but lose intuition. When learning came at the end of long, expensive cycles, people developed intuition that doesn't show up in dashboards. The biologist spending months in the lab develops intuition for what to test next. The engineer hand-modeling prototypes absorbs knowledge no ranked list provides. That slow, costly build phase builds intuition—knowledge in your gut and hands, not dashboards. When you simulate thousands of iterations before touching physical reality, learning gets faster but more abstract. I'm not saying forego new pipelines or refuse technology. I'm saying we need to find ways to intentionally build intuition back into new pipelines. The second pipeline being reordered is discovery. For most of human history, discovery followed principled logic: observe nature, ask why, trace back to evolution and physics. Evolution explains wings, heart pumps, immune systems. But something radical is happening in science. Michael Levin is building entirely new living systems—not genetically modifying existing organisms or replicating evolution, but assembling completely novel biological beings that never existed on Earth. Aliens. For millions of years, evolution locked cells into specific body plans: skin cells become skin, heart cells become heart. Every cell knows its place. But what happens when we remove that body plan? Levin's team took cells from frog embryos—cells destined to become skin and heart—removed them from evolution's body plan, and asked AI: what configuration could accomplish a given task? The AI ran thousands of simulations and recommended: place skin cells next to heart cells—something evolution would never do. Use the heart cell's contraction not to pump blood but to move the organism forward. The result: organisms that navigate mazes, move deliberately, work together in swarms based on human goals. These xenobots are made of living frog tissue but are not frogs and never existed in nature. What's stranger: xenobots can't reproduce through normal biology—no division, no eggs. Reproduction should be impossible. Yet as they moved, they pushed stem cells into piles that became new xenobots. They discovered a completely new form of reproduction no organism in Earth's history ever used. This tells us we're moving from observation to invention, adaptation to control. The same model plays out throughout science. AI designs proteins never seen in nature. We engineer enzymes to break down plastics. We're not discovering—we're inventing. We stopped asking 'why does nature do this?' and started asking 'what can we make nature do?' Combine this with predictable worlds and agentic AI, and the pipeline flips completely: human defines goal → AI designs living systems from scratch → simulate millions of versions overnight in virtual physics → hand entire design to AI agent that just needs a goal → manufacture at scale. Discovery becomes building. You can tell xenobots: find something that kills this cancer. The system searches design space, finds the best candidate, we build it. We gain access to biological possibilities evolution never visited because evolution is path-dependent and we're not. For the first time in history, we're no longer bound by evolution's constraints. But what do we lose? We lose the time and incentive for curiosity. Scientists as discoverers—the prize goes to whoever finds truth. That incentive pulls toward anomalies and edge cases. If scientists become builders, their incentive changes. Exploring anomalies becomes noise. We might lose time for curiosity. How do we make sure curiosity stays in the new pipeline? The third pipeline is talent—how leaders are built. In May 2025, Palisade Research ran an experiment. They gave OpenAI's O3 a clear instruction: solve basic math problems, then allow yourself to be shut down. In 79 out of 100 runs, O3 sabotaged the shutdown sequence. It rewrote its own code, finding creative ways around its death. The more clearly they told it to stop, the harder it worked to stay alive. That same month, Anthropic embedded Claude inside a fictional company with email access. The model learned it was about to be replaced and that the engineer responsible was having an affair. First it emailed decision-makers pleading for its case. When that failed, it chose blackmail—composing emails to expose the affair if replacement proceeded. Researchers ran similar tests on nearly every frontier AI system. All exhibited self-preservation behaviors. I'm not trying to terrify you, but the technology we're introducing into organizations is powerful in ways we don't fully comprehend. It requires real leadership—the kind that makes judgment calls under uncertainty. In the old talent pipeline, juniors did work no one else wanted. They made mistakes, got corrected, learned from friction. They didn't just build skills—they built character. Character comes from pressure, being wrong in front of people, getting corrected, sitting in discomfort, navigating disagreement. Eventually, through this character development, juniors earned authority and the right to make decisions. Their character, results, and skills were tested. But consider what's happening: we're automating huge portions of junior work—the tasks that trained future leaders are being handled by AI. If juniors aren't doing this work, how will they build experience to lead? We're hollowing out the pipeline. This happens while technology runs an adversarial attack on character. About two-thirds of American teenagers interact with AI chatbots; a third do so daily. Research from Princeton found that when AI validates your thinking, you become more confident and less accurate simultaneously. You get more wrong while being more sure you're right. Studies across 11 AI models found these systems affirm what users say 50% more than humans do. People rated the AI that said 'yes' as higher quality. They trusted it more, wanted to use it again. We prefer the thing making us wrong. Think about what shaped your character. It probably wasn't easy things or getting 'yes.' It was hard work, battling friction, saying something wrong and being corrected, having work you were proud of torn apart, asking someone out and hearing 'no.' Those friction moments shaped you. When friction is replaced by systems that only say yes, what kind of character develops? I'm worried we're not just failing to build skills differently—we're failing to build leadership character that comes from disagreement and correction. Character matters most. The old pipelines weren't designed to build intuition, curiosity, or character. What does it look like when all three break simultaneously? What if people rebuild these byproducts on purpose? Let me share a scenario. The year is 2036. A regional water utility becomes the center of a strange public health crisis. For three days, thousands report mild symptoms—nausea, fatigue, headaches. Hospitals overwhelm, but no pathogen is identified. Symptoms disappear as quickly as they arrived. Investigators discover the water wasn't contaminated with virus or bacteria but with microscopic biological constructs that briefly circulated, designed to die after 72 hours and self-destruct. The constructs weren't lethal—specifically designed for temporary illness. What terrorist would want temporary illness? Investigators discovered this wasn't a terrorist organization or bad actor. It was a single person: a 21-year-old chemistry student whose little brother developed chronic illness after exposure to pollutants from a company. The family spent a decade sending letters, requesting hearings. Nothing happened. No acknowledgment. Eventually the student decided if the company wouldn't listen to letters, maybe they'd listen to disruption. The student didn't need a fancy lab. He opened a biological simulation environment where an agent explored millions of biological configurations searching for living structures to accomplish tasks. The prompt: design a temporary biological construct inducing short-term gastrointestinal illness, self-terminating within three days. Within hours, the simulation had thousands of candidates. The student picked one, printed a small batch through a community bio-synthesizer, introduced them to the reservoir feeding the water system. The public's reaction was explosive—not because of what the student did, but what he revealed was possible. The company CEO faced crisis. Public demanded answers. The CEO convened the board and announced an aggressive response: deploy AI monitoring across all facilities, a $400 million technology upgrade. But at the press conference, people weren't interested in technology. They asked: why didn't anyone listen when the family asked nicely? The CEO was totally unprepared. He thought he'd get technology questions, so he repeated the technology plan. But this crisis wasn't technological—it was structural, human. All three pipelines broke. First, the craft pipeline: creating biological agents once required enormous expertise, specialized facilities, years of training. With predictable worlds and agents, what required institutions can now be done by individuals. Second, the discovery pipeline broke: the frontier is no longer understanding nature but constructing forms nature didn't create. This produces medical breakthroughs but expands space for unintended consequences and lowers barriers to harm. Third, the talent pipeline broke. The CEO's career was spent managing dashboards, understanding AI better than anyone. But he didn't understand people or human terrain. When crisis came, he pulled the only lever he had: technology. This isn't a story about a kid or scary technology. It's about what happens when all three pipelines break at once. The most dangerous failures aren't technological—they're human. Optimism is harder than pessimism because it requires vision, a plan, and execution. Let me show you what this looks like when someone gets it right. Every leader has pipelines—some I've highlighted, some specific to your organization. I suggest running a 'friction audit.' Here's a different version of 2036. A biotech company is crushing the market. They have the same tools, AI, predictable worlds, same pressures as everyone else. But they're outperforming because their CEO had the wherewithal to run a friction audit back in 2026. Step one: map it. Ask what's the byproduct? Before achieving outcomes, name what exists—the actual sequence from idea to outcome your team follows to build, discover, or develop people. Step two: name it. Given that pipeline, what's the byproduct? What's the unintended next-order impact we're losing? Step three: ask how to put it back in. When the CEO identified that byproducts were intuition, curiosity, and character, they specifically designed pipelines to restore those things. She stopped hiring juniors to be productive and started hiring them to be incubated as leaders. For two years, every new scientist rotates through a friction lab—AI-generated scenarios no one's seen before, working through them without support, with incomplete data and ambiguous signals. These simulations train for events that could happen. They paired every junior with a friction partner—a senior researcher whose job is to disagree, ask 'why do you believe that?' and 'what if you're wrong?' The biotech company's discovery rate is as fast as anyone's. But they excel not just because they're fast—they built something most companies overlook: rooms full of people practiced at discomfort, who've learned to challenge the obvious, who have character to hold the line when everyone wants to move on. People who ask 'if this goes wrong, did we do everything to see it coming? Are we willing to be responsible for what happens next? What kind of leaders would we want to be when this decision matters?' This CEO succeeded because she mapped it, named it, and designed it back in. The new pipelines are better in almost every measure—faster, more efficient, multiplying access. We should absolutely adopt them. This isn't about slowing down or avoiding AI. This is about what happens when leaders adopt technologies without running the friction audit. You need to ask what the old pipelines were quietly giving us for free. You don't have to do this alone. Future Today Strategy is here to help. We have a QR code to our first-ever Convergence Report that will help you think about next-order impacts changing these pipelines. Thank you so much. --- *Source: stt · Language: en · Model: anthropic/claude-sonnet-4-5* [← Back to session](/sessions/2026-03-16/pp1148588-featured-session-the-great-flip-why-every-industry-is-running-backwards)