Hey Reader,

Welcome to the 25th edition of the 3-2-1 (check out previous issues here).

Trying something new today and starting with 1 Idea From Me before dropping the interesting links for this week. Let's get into it.

---

1 Idea from Me

Your inbox is a destination group. You need a journey group.

The head of L&D at a 3,300-person tech company told me last week that she tunes out most emails, LinkedIn posts, and webinar invites, but for a reason that surprised me.

She's in the middle of some of the hardest work in our field: figuring out what AI should change in how work gets done, how to segment her learner population to make people AI-productive, and how to do all this with a lean team serving people who range from "never opened Claude" to "blowing it out of the water."

She has lots of questions, but she doesn't tune out because the content isn't helpful or there's too much. She tunes out because it's too finished and polished. She's not ready for the final version because she's 5 steps away from it.

"One day I'll come back to you and be like, we have these beautifully segmented programs, and they have clear audiences and objectives, but... I'm not there yet, and a lot of what I receive is positioned as being there."

Then she told me where she actually gets help. Two former colleagues at different orgs that she can call anytime. They've seen similar initiatives, can tell her what worked and what didn't, and help her think through how to apply the lessons they learned.

That's thought partnership, and it's the one thing knowledge work doesn't have enough of.

Two kinds of groups

In 2020 I wrote that transformational learning needs two kinds of groups. Destination groups are people already where you want to be. They hand you frameworks, models, and proof it's possible. Journey groups are peers at your stage, working on the same problem at the same time. Destination groups set direction, but journey groups get you there.

That head of L&D's inbox is one enormous destination group: finished frameworks, concise case studies, perfect playbooks. The results are always in the headline. None help with the very specific next question she's pondering.

Her journey group is those two former colleagues. She's lucky she has them. Many talent leaders I know are working on lean teams or in silos. They're usually the only one owning strategy for their area. No one to bounce ideas off.

We all need someone to bounce ideas off, even one of the greatest writers in history.

No Lord of the Rings without this guy

J.R.R. Tolkien read The Lord of the Rings aloud to C.S. Lewis chapter by chapter as he wrote it, over more than a decade. Lewis had his own unfinished books at the time, but he showed up every week to hear the next chapter and ask for more.

"But for his interest... I should never have brought The Lord of the Rings to a conclusion." - Tolkien

The power of thought partnership is well studied in science. Dunbar (first link below) shows that scientific breakthroughs happened when researchers put confusing, half-done results in front of peers who were running their own confusing experiments down the hall.

We've been watching this for two years working on judgment with clients.

Building judgment by pressure test

A senior finance leader in one of our programs spent this year leading the renegotiation of a major credit facility across five functions with conflicting goals. Legal wanted minimal change. Accounting wanted nothing restated. Her CFO wanted better terms and a bigger line of credit.

Throughout the negotiation, she ran her approach by her CFO before joining any meeting. By doing so, she caught things that mattered to a function she had never worked in, so when she was in the meeting, she was prepared and speaking their language.

We've codified this as a behavior: pressure-test before the room. The question that triggers it is one I use all the time: "who can tell me if I'm wrong the quickest?" This question short-circuits our need to be right, which is the main thing that gets in the way of getting to the right outcome.

The AI catch

I use AI as a thought partner every day, but you have to use it carefully.

AI was trained on every finished product ever published. Left to itself, it pulls you toward the polish the L&D leader I opened with was avoiding. Accepting the final version AI gives you is cognitive surrender, letting the machine make the call and adopting it as your own.

That's why human thought partnership keeps its value: a peer at step 3 knows what step 3 feels like and can help you navigate the details, while you own the final call.

Your move this week

Even if you already have go-to peers who help you organize your messy thinking, find someone new to do that with this week. Ask yourself, "who can tell me if I'm wrong the quickest?" and see who comes to mind. Fresh perspectives can expose your best thinking.

If you want help building this into how your team works, hit reply.

---

3 Things for Work (in L&D)

Where Good Ideas Come From (Steven Johnson, TED)

• Johnson tells the story of Kevin Dunbar, who put cameras in molecular biology labs to find out where breakthrough ideas came from. Almost none came from a scientist alone at the microscope. They came from the weekly lab meeting, when people put unfinished data and the noise in their signal in front of peers.

• 🏋 Effort ≈ 18 min watch

How Pixar Fosters Collective Creativity (HBR, Ed Catmull)

• Pixar operationalizes thought partnership in two ways. Dailies, where animators show unfinished work to the whole team every day, because "once people get over the embarrassment of showing work still in progress, they become more creative." And the brain trust, master filmmakers who give candid notes on a film in progress and hold no authority over what the director does with them.

• 🏋 Effort ≈ 15 min read (from 2008, and it has aged well)

Meaningful Long-Term Thought Partnerships of Minds and Machines (Current Directions in Psychological Science, 2026)

• Cognitive scientists ask what it would take for an AI to deserve the name thought partner over months, rather than one chat. Their answer is the human recipe: build a model of the other person, establish common ground, make joint plans, and keep doing it over time.

• 🏋 Effort ≈ 20 min read (8 pages; the first 2 carry it)

---

2 Things for Life

How Two Trailblazing Psychologists Turned the World of Decision Science Upside Down (Michael Lewis, Vanity Fair)

• Daniel Kahneman and Amos Tversky flipped a coin for first authorship on their first paper and alternated from then on, because neither could say whose idea was whose. Kahneman later wrote that they came to know each other's minds "almost as well as our own." Their subject was judgment, and the best part is watching two people build it together.

• 🏋 Effort ≈ 25 min read

A line from Russ Roberts, in My Twelve Rules for Life:

"Learn to enjoy saying 'I don't know'... When you're young, it hurts to say it. As you get older, you can savor saying it... Remember that honesty is better than posing as the expert that you're not."

• Asking who can tell you you're wrong the quickest only works if you can say these three words first. It gets easier with age, and the people around you trust you more for it.

---

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

I'll be back in two weeks ✌️

Andrew

P.S. I'm testing an idea with a partner. A journey group for senior talent leaders. 8 weeks, 12 hand-picked people, each bringing one live initiative they're measured on. Some sessions are the group working on your initiative. The rest are building AI workflows you keep. Interested? Hit reply and let me know. Cooking up an early bird offer for subscribers.