RESHUFFLE An interactive companion to the book
Explainer · Chapter 2 ▸ Frames cluster

“How smart is AI?” is the wrong question.

Ask what the system can now do.

Most leadership teams are debating whether AI is smart enough yet. That debate is the wrong watch-list. While they're tracking model capability, the system around their business is being restructured by competitors who are reading AI through a different lens entirely.

The frame you bring to AI decides what you can see - and most leaders are looking through the wrong one.

▍ The mechanism

Two frames, two watch-lists

Public debate fixates on a single question: how smart is AI? Can it pass the bar, write the novel, hold the conversation. Even when the answer is yes, this frame leads to the wrong actions.

The frame matters because it tells you what to watch. The IQ frame says watch model capability - GPT-4 vs GPT-5, Claude vs the next thing. The system frame says watch which constraint in your workflow just stopped being scarce. Two completely different watch-lists, leading to two completely different strategies.

Whoever frames the question chooses the answer. The IQ frame leads to “let's wait until the model is smart enough.” The system frame leads to “what just got coordinated that couldn't be coordinated before, and who owns that coordination layer?” Same technology, different frames, different winners.

▍ Historical analogue

The Maginot Line

A perfect answer to the wrong question - 1940 edition.

Through the 1930s, traumatised by the trenches of the Great War, France built the Maginot Line: a 450-mile chain of fortified bunkers, turrets, and underground rail along its eastern border with Germany. By every technical measure, it was a masterpiece. Unbreachable. Heavily reinforced. The perfect defense.

In May 1940, German forces simply went around it - through the Ardennes, which the French had considered impassable. What they brought wasn't a better fort or a smarter weapon. They brought a new system of warfare: mechanised infantry, tank columns, and dive bombers all coordinated in real time via two-way radio. Blitzkrieg.

The Maginot Line stood through the entire attack. Unbreached. And utterly useless. It had succeeded as an engineering solution and failed as a strategic response, because while France had perfected an answer to centralized, linear warfare, Germany had changed the system of warfare itself.

Most companies are building Maginot Lines around their AI adoption right now. Bigger models. Smarter tools. Faster inference. Meanwhile, the system of work they sit inside is being rewired by competitors playing a different game.

▍ Two lenses on the same thing

What each frame sees

Frame A
The IQ frame
Watches
Model capability. Benchmarks. Vendor announcements.
Asks
“Is the model smart enough yet?”
Plans
Wait for capability to mature. Pilot the new tool. Roll out widely.
Action
Buy a smarter tool. Train staff to use it.
Outcome
Productivity bump on existing tasks. Margins compress as everyone runs the same play.
Frame B
The system frame
Watches
Which constraint just dissolved. Where coordination is now possible.
Asks
“What can the system now do that it couldn't before?”
Plans
Re-architect workflows. Claim the new coordination layer. Reposition the business.
Action
Rewire the system around the new architecture.
Outcome
New basis of competition. New control points. Industry-level repositioning.
▍ How this plays out

Two examples

01

The typist and the word processor

The typist's job was a bundle of typing, formatting, error correction, and revision. The role existed because each of those tasks was expensive - fixing a single error meant retyping a page; revisions meant manually managing versions across a typing pool.

The word processor arrived. Through the IQ frame, it looked like a productivity tool that would make typists faster. Through the system frame, it was the death of the role: the costs that justified the typist bundle had just collapsed to zero. Typing didn't disappear - it became a distributed task that every knowledge worker absorbed into their own work.

No amount of reskilling would have saved the typist. The system no longer needed a role organised around that task. Companies that watched typist productivity missed it. Companies that watched the workflow restructuring saw it coming years out.

02

Amazon's Kiva robots

Amazon's Kiva robots roll through warehouses bringing shelves of products to stationary pickers. The IQ frame reads this as “robots assist warehouse workers with picking and packing.” Task-level automation. Productivity gain.

The system frame reads it as the warehouse becoming one coordinated node in a global same-day-delivery system. The robots aren't there to help the worker. They're there to let Amazon redesign the architecture of fulfillment - what gets stocked where, how returns flow back, how the warehouse talks to last-mile delivery and to customer preference data.

Worker productivity is a downstream effect of a much larger system restructuring. Anyone watching only the picker–robot interaction was reading a different book than the one Amazon was writing.

▍ Apply it

Which frame are you running?

Five quick questions. Pick the answer that sounds more like you. Honesty beats aspiration - the value is in finding out where you are.

  1. 01

    When evaluating a new AI investment, your team's first instinct is to compare:

  2. 02

    Your AI watch-list is mostly tracking:

  3. 03

    When a competitor launches an AI feature you don't have, your response is to:

  4. 04

    Your strongest concern about AI displacement is:

  5. 05

    If you had to bet on who wins the AI era, it's: