Living edition19 theses under observation
The state of the theses
A book about exponential change has to take its own medicine. Here are its key theses, checked against what has happened since print — with the linked evidence behind every status.
7 Reinforced3 Nuanced0 Challenged9 Holding
The four states
- ReinforcedRecent evidence actively supports it.
- NuancedEvidence adds real context: read it with that nuance.
- ChallengedThere is evidence against it. We'll say it here, not hide it.
- HoldingHolds exactly as printed.
How it's audited: every chapter has verified news and data linked in the editorial CMS (typed: supports, challenges, updates). The editorial crawler adds signals weekly; each thesis status is reviewed by the editors against that evidence. Any status other than “holding” requires linked evidence.
Chapter 01The Exponential Moment
Read the chapter →ReinforcedThe exponential gap is operational: it separates what technology already enables from what companies, careers and societies can absorb.
Mass adoption without returns confirms the gap: almost every company uses AI, very few have turned it into results.
- McKinsey: el 88% usa IA en alguna función, pero solo un tercio escala — y el EBIT impactado es <5% (2025-11)
Reviewed · 2026-06-10
Chapter 02The Three T’s Paradigm
Read the chapter →Chapter 03The Three Forces of Change and Where They Lead
Read the chapter →Chapter 04It Is Already Happening
Read the chapter →ReinforcedThe right question is no longer whether AI will change your sector: it is whether you will be part of that transformation.
Pilots have become deployments: robots building cars at scale and software agents valued in the tens of billions.
- El robot Figure 02 ayuda a fabricar 30.000 BMW X3 en Spartanburg tras un piloto de 11 meses (2025-10)
- Boston Dynamics presenta el Atlas comercial en CES 2026 — primeras flotas para Hyundai y DeepMind (2026-01)
- Cognition (Devin) negocia ronda a 25.000 M$ — multiplica por 6 su valoración en seis meses (2026-04)
Reviewed · 2026-06-10
Chapter 05Time Horizons and How to Navigate Them
Read the chapter →Chapter 06ExOS, The Exponential Operating System, a Compass Before the Journey
Read the chapter →Chapter 07Automate, When What You Used to Do Now Happens by Itself
Read the chapter →ReinforcedAutomation is the first answer: the same as always, faster and cheaper. Real disruption lives in the second one: using AI to do what wasn't possible before.
The second answer now ships as products: agents that use a computer like a human, and humanoids leaving the pilot factory.
- Anthropic abre "Computer Use" — Claude mueve cursor, hace clic y escribe como un humano (2024-10)
- OpenAI lanza Operator — el primer agente CUA para usuarios Pro (2025-01)
- BMW lleva los humanoides a Europa — piloto AEON en Leipzig y nuevo "Center of Competence" (2026-04)
Reviewed · 2026-06-10
Chapter 08Augment, Amplified Human Intelligence
Read the chapter →ReinforcedAI amplifies what you already are. The next great differentiator is not knowing AI: it is knowing what to do when AI no longer differentiates you.
The market already prices amplification: a double-digit wage premium for AI skills, and half of jobs delegating substantial tasks to assistants.
- Prima salarial por habilidades IA: 56% (PwC Global AI Jobs Barometer, observatorio) (2026-05)
- Empleos con ≥25% de tareas hechas con Claude: ~49% (Anthropic Economic Index, observatorio) (2026-05)
Reviewed · 2026-06-10
Chapter 09Anticipate, From Reacting to Getting Ahead
Read the chapter →Chapter 10The Four Redefinitions — Creating, Working, Learning, Governing
Read the chapter →ReinforcedWhat AI redefines is not what you do: it is what being human means across the four dimensions where we build ourselves — creating, working, learning and governing.
All four dimensions already show measurable signals: generative creation open to the public, and AI tutors scaling an order of magnitude in one school year.
- Khanmigo crece de 68.000 a 700.000 estudiantes en un curso escolar (2025-09)
- OpenAI abre Sora al público — texto a vídeo en alta definición para usuarios Plus y Pro (2024-12)
Reviewed · 2026-06-10
Chapter 11Society Ahead of Institutions
Read the chapter →NuancedCitizens adopt technology before institutions know how to name it: society runs ahead of those who govern it.
Institutions have started to catch up: the AI Act is now phasing into enforcement and courts are imposing record settlements. The thesis still holds on speed, but it no longer describes a vacuum.
- Entran en vigor las obligaciones de gobernanza y modelos de uso general del AI Act (2025-08)
- Anthropic acuerda 1.500 M$ con autores por entrenamiento sobre libros pirateados — récord histórico (2025-09)
Reviewed · 2026-06-10
Chapter 12The Adaptation Crisis
Read the chapter →NuancedThe adaptation crisis is not solved with more information: it is solved by redesigning the structures that prevent acting on the information we already have.
Macro evidence hasn't settled the question yet: no aggregate productivity impact, but 30% gains where work was actually redesigned. Both readings fit — time will tell which weighs more.
- Goldman Sachs no ve impacto agregado de la IA en productividad — pero sí 30% en dos casos concretos (2026-03)
Reviewed · 2026-06-10
Chapter 13The Technological Gap and Our Capacity to Respond
Read the chapter →ReinforcedThe technology gap is not closed by buying technology: it is closed by building the capacity to absorb it.
The data says it alone: near-universal usage, a third scaling, still-marginal impact on results. Buying is not absorbing.
- McKinsey: el 88% usa IA en alguna función, pero solo un tercio escala — y el EBIT impactado es <5% (2025-11)
Reviewed · 2026-06-10
Chapter 14Talent and Resilience, The Skills Race
Read the chapter →ReinforcedThe race is not against the machine: it is for skills. Most of the workforce will need reskilling before 2030.
All three observatory series point the same way: mass reskilling, shifting skills, and a growing wage premium for AI abilities.
- Trabajadores que necesitarán reskilling/upskilling: 59 de cada 100 (WEF, observatorio) (2026-05)
- Skills que cambiarán a 2030: 70% (LinkedIn Economic Graph, observatorio) (2026-05)
Reviewed · 2026-06-10
Chapter 15The Challenge of Trust and Ethics
Read the chapter →NuancedThe ethical question about AI is not what it can do: it is who answers when it does. And that question has no technical answer — only a political one.
The political answer exists, but diverges by bloc: deregulation in the US (EO 14179) while the European AI Act enters its enforceable, finable phase. The thesis is right that it's political; convergence, for now, isn't coming.
- Trump rescinde la orden ejecutiva de Biden sobre IA y firma EO 14179 "Removing Barriers to American AI" (2025-01)
- Cuenta atrás para 2 de agosto de 2026 — el grueso del AI Act se vuelve aplicable y multable (2025-11)
Reviewed · 2026-06-10