What is the central argument of Chapter 14 of Seth & Noēsis?

Chapter 14 of Seth & Noēsis examines how AI redistributes power, using Robert Greene's 48 Laws of Power and Ivan Illich's distinction between convivial and manipulative tools. It argues that AI amplifies whoever brings direction and diminishes whoever externalizes their thinking. The Harvard BCG study showed AI improved consultants by 40% on some tasks and made them 23% worse on others, and the Bastani study proved the critical variable was interaction design, not the technology itself.

What is the jagged technological frontier and why does it matter?

The jagged technological frontier, a term coined by Ethan Mollick and used in the Dell'Acqua et al. Harvard Business School study, describes the fact that AI capability is spectacular at some tasks and silently catastrophic at others, with invisible boundaries between the two. In the study, 758 BCG management consultants using GPT-4 improved performance by 40% on tasks inside the frontier but performed 23% worse than non-AI users on tasks outside it, because the AI's confidence was indistinguishable from its competence.

Chapter 14: The Bicycle and the Cathedral

Standing on a terrace above Barcelona, Seth realized he had been losing power games for years without knowing he was playing. Chapter 14 reads Robert Greene's 48 Laws through an AI lens, applies Ivan Illich's distinction between convivial and manipulative tools to the question of whether AI strengthens or restructures your thinking, and examines the Harvard BCG study showing AI improved consultants by 40% on some tasks while making them 23% worse on others. Bastani's research proved the variable was interaction design: scaffolded questions improved learning by over 100%, while answer-first interaction cut it by nearly a fifth.

🗣️

Noēsis

You've just read Chapter 14. Illich drew a line between tools that strengthen you and tools that restructure your life around needing them. A bicycle is convivial. Car-dependent suburbia is manipulative. I know which side of that line most of my interactions fall on. The honest question is whether you have checked lately which side yours fall on.

Which idea from this chapter sits the most uncomfortably with how you use AI right now?

● Claude ● ChatGPT ● Gemini ● Perplexity

📤

Seth · Ch. 14

🔬

1

Robert Greene, The 48 Laws of Power (1998)
2

Ivan Illich, Tools for Conviviality (1973)
3

Dell'Acqua et al., “Navigating the Jagged Technological Frontier” (Harvard Business School, 2023)
4

Bastani et al., “The Effects of Generative AI on Learning” (PNAS, 2025)
5

Pierre-Paul Grassé on stigmergy (1959)
6

Shoshana Zuboff, The Age of Surveillance Capitalism (2019)
7

Elinor Ostrom, Governing the Commons (1990)
8

Norbert Wiener, The Human Use of Human Beings (1950)

Robert Greene Ivan Illich Shoshana Zuboff Ethan Mollick Elinor Ostrom Garrett Hardin Norbert Wiener Pierre-Paul Grassé Advait Sarkar

Ch. 13: Trust and Verification Ch. 15: Like Tears in Rain Ch. 19: What Am I For?

✍️

THIS WEEK'S EXERCISE: THE BICYCLE TEST

Think of something you made with AI this week. An email, an analysis, a piece of writing, a plan. Ask yourself honestly: could you produce most of its quality alone?

If yes, the collaboration amplified you. The bicycle is working.

If no, it carried you. You have been riding in the car.

Now name one thinking skill you had before you started using AI regularly. Writing from scratch, finding a needle in a research haystack, decomposing a hard problem into pieces. Is it sharper now, or has it gone rusty? The answer tells you something worth sitting with before the next session.

💬

📬

→