Hey Reader,

Welcome to the 23rd edition of the 3-2-1 (check out previous issues here).

This week I sat down with alumni of a thinking-skills program we've been running for almost two years. One of them, now leading a team, described working with AI as a pressure cooker. Everyone is expected to have an answer instantly, and AI makes instant answers easy. That conversation is the theme of 1 Idea From Me below, plus I've got two new papers that landed on opposite sides of the same argument up first.

Let's get into it.

---

3 Things for Work (in L&D)

Five principles for designing brain-powered organizations (McKinsey)

• McKinsey now argues that the cognitive capacity of your people is a strategic asset to be designed for, not a personal wellness perk. If thinking is an asset, then how your organization protects it, develops it, and spends it becomes a leadership decision, not an individual one.
• 🏋 Effort ≈ 5 min read

Thinking Fast, Slow, Artificially: AI and Your Brain (Wharton)

• The flip side. A smart learning team shared this with me. Wharton researchers extend Kahneman's two systems of thinking with a third: artificial cognition that lives outside your brain. Their warning is a term you'll be hearing a lot: cognitive surrender. Across experiments with 1,300+ people, the more participants trusted AI, the more likely they were to follow its incorrect advice without questioning it.
• 🏋 Effort ≈ 7 min read

How to Win Buy-In for Your L&D Strategy Using AI (my lightning lesson on Maven)

• This is me, live on Tuesday, August 18 at 1pm ET, showing how to use AI to enhance your thinking while preserving it. Sign up even if you can't attend live. I'll send you the recording and the exact prompt we use (it also helps me as the algo ranks on total registrations🙏).
• 🏋 Effort ≈ 30s to register

---

2 Things for Life

Is this Norwegian cow scratching its back, or twerking? (X)

• Watch it and pick a side. I've watched it five times and I'm still not sure.
• 🏋 ~1 min watch

A line from E.O. Wilson, written in 1998:

"We are drowning in information, while starving for wisdom. The world henceforth will be run by synthesizers, people able to put together the right information at the right time, think critically about it, and make important choices wisely."

AI didn't make this less true. It made it the job description. Which brings me to this week's idea.

---

1 Idea from Me

Don't surrender your reps

The two papers I shared above belong in the same conversation.

McKinsey says the cognitive capacity of the people inside your company is becoming a strategic asset. Wharton says we're handing that asset away, one AI-drafted decision at a time. They coined it cognitive surrender. Not offloading a task, but letting the machine make the call and adopting it as your own.

The tension between the papers is the defining people question of the next few years: how do you use AI aggressively and still develop the judgment of your people?

What that tension looks like on the ground.

This week I spoke with alumni of a thinking-skills program we've been running for almost two years. One participant described the global push to amplify everything with AI as a pressure cooker, and the people feeling it most are product managers, designers, and engineers. Overnight, they're expected to demonstrate rationale, hypothesis quality, and the ability to negotiate and communicate across functions, while roles and ownership boundaries shift under their feet.

She participated in an early cohort and has since been promoted to lead a team: "I'm trying to teach my team how to operate in those different modes instantly," she told me. "And that's hard."

The environment rewards immediate answers. But some people need time to process, and she's coaching them to claim it: "If you need time to think, say you need time to think and you will come back with an answer."

That line is deceptively radical. In a room that rewards instant answers, saying "I need time to think" is a statement that your thinking is worth waiting for. Most people would rather bluff, or reach for the machine.

That pressure is what makes surrender attractive. AI will hand you the research, the synthesis, the hypothesis, the design decision, right now. Her worry is the same one Wharton published data on: "I'm a little afraid that we're going to be over-reliant on AI, because you just need reps."

You just need reps. That's the whole argument in four words.

Taste and judgment are not downloaded. They're built the same way strength is: through repetitions under load.

McKinsey's strategic asset, surrendered.

Every research synthesis you do yourself, every hypothesis you form and test against reality, every recommendation you defend in a room, is a rep. Automate all of them and the work still ships. But what stops is the development of the people shipping it. You end up with a company that has answers everywhere and judgment nowhere.

How to build judgment.

After almost two years of doing this work, here's what we've learned actually builds judgment. Three conditions are necessary:

1. Practice on live work. Not simulations, not case studies from someone else's company. The problem on your plate this week, written down and worked out loud. Judgment is contextual, so it has to be built in context.

2. Practice with peers across the business. Judgment develops fastest when your reasoning collides with perspectives you don't have: the engineer hears how finance sees it, the designer hears what legal is afraid of. Peers are the resistance that makes the rep count.

3. Limit AI from giving you the answer. Use it for everything except the final recommendation. That's your point of view, you should own it. We've been building an AI companion for our clients designed to do exactly this: it questions, challenges, and sharpens your thinking, but refuses to hand you the conclusion.

Thinking critically to arrive at a point of view is what every company is now asking of knowledge workers. It's messy, uncomfortable, and slow. You can't get around that. You can create the space and the psychological safety for it to happen, and then repeat the loop quickly and repeatedly.

Do that and people develop judgment, which is what drives business results.

Here's your move this week. Find one decision you've been letting AI finish, and run it as a rep instead: AI for the research, you out loud for the call.

---

That's it for this week - enjoy your Sunday!

I'll be back in two weeks ✌️

Andrew

P.S. 👉 If the pressure cooker sounds like your company, the lightning lesson above is where I'll show you the working alternative. Register here and you'll get the recording and the prompt even if you can't make it live.