> ## Content Index
> Fetch the complete content index at: https://www.cpj.fyi/llms.txt
> Use this file to discover other available public pages before exploring further.

# Podcast: Self-organization as recursive self-improvement
- URL: https://www.cpj.fyi/radar/podcast-self-organization-as-recursive-self-improvement/
- Published: 2026-11-17T17:22:40.000Z
- Updated: 2026-11-17T17:22:40.000Z
- Description: Peter Merholz, Jesse James Garrett, and I talk about role clarity, structural safety, and getting teams unstuck.
- Author: Clay Parker Jones
- Tags: Radar, Speaking, Org Design, Artificial Intelligence

I joined Peter and Jesse on [Finding Our Way](https://podcasts.apple.com/us/podcast/finding-our-way/id1510667674?i=1000797944976&ref=cpj.fyi). I’ve looked up to them since their Adaptive Path days. (I wanted to be a UX designer when I was younger!)

The conversation grew out of [The End of Role Clarity](https://www.cpj.fyi/essays/the-end-of-role-clarity/). Peter had some quite reasonable objections: roles help people understand work, give organizations something to administer, and carry identities people care about. And not for nothing, one of the things I loved about this episode is that it's not just me answering questions, but an *actual* conversation about the topic from both ends.

A few favorite bits, bobs, what-have-yous:

## *Clarity at the team level*

> We talk about, "AI is going to be this enabler of [super performance](https://www.cpj.fyi/essays/super-performance/) for teams." We already have it. It’s just a good team. **A good team is super intelligent.** We’ve all experienced, in the best teams that we’ve worked on, this feeling of a hive mind. That’s what we’re going for here. \[…\] And what I find is, over and over, the teams that actually spend a bit of time getting their shit together up front, they end up delivering.

This is what a [Team Charter](https://www.cpj.fyi/43-team-charter/) is for. Write it with the people who will use it. Include the work you’ve agreed to leave out. Test it against a recent disagreement: would it have helped you make the decision? Revise it when the work changes. A charter is worth the time it takes to maintain.

Two additional points here:

1. AI is also enabling [super performance](https://www.cpj.fyi/essays/super-performance/)
2. Self-organization is recursive self-improvement (RSI) for people

## *Safety needs rules*

> We have to have, yes, good team kickoffs. We need to ask good questions. We need to model, as leaders, vulnerability. We also need to address the structural parts. \[…\] We need to address the lack of **structural safety** within an organization, because you’ll have leaders within an organization that want to build psychological safety on their team, but they are not able to because that team could be deleted at the drop of a hat.

[Structural and Psychological Safety](https://www.cpj.fyi/3-structural-and-psychological-safety/) and [Rule of Law](https://www.cpj.fyi/2-rule-of-law/) sit near the beginning of the book for this reason: they are foundational to the rest of the stuff working, and they are things that you have to build toward. People need to know what authority they hold and whether the rules apply to their boss, too. An invitation to take initiative becomes a very different proposition when someone senior can casually revoke the permission.

Those protections should also make change understandable. Keeping a team safe shouldn't have to mean preserving its current shape forever.

## *Find the willing team*

> Don’t convince people. Don’t try. Just offer the thing, offer the better way of doing something, offer the better design, offer the better practice, and if they don’t wanna do it, they probably have a good reason, and trying to trick them into doing it is not worth your time. Instead, a better plan is to go find teams that really want to get something done, lavish your time and attention and tooling on them, and then show the organization what they’re able to do.

The [Demos pattern](https://www.cpj.fyi/35-demos/) is useful here, and while it's a pretty straightforward idea (let people try the work!), I really like this book chapter because it's not always easy for everyone to do. Here's a lil' excerpt.

| Team                        | What & How to Demo                                                                                                      | Notes                                                                                                |
| --------------------------- | ----------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------- |
| Product, Engineering *Easy* | Working builds, functional prototypes, actual systems Live interactive sessions showing real functionality              | Focus on what works now, not what will work eventually; use real data; demonstrate failure modes.    |
| Design *Easy*               | Interactive prototypes, user journeys, visual systems Guided experiences with actual user interactions                  | Show design in context of use; demonstrate multiple states; reveal the system beyond single screens. |
| Marketing *Medium*          | Actual campaigns, real content, real channels Live channel reviews, real audience interactions                          | Show performance data; demonstrate targeting mechanisms; reveal the full customer journey.           |
| Facilities, Space *Medium*  | Physical environments, actual usage, environmental metrics Guided tours, video of actual use, demonstration of features | Show spaces being used; demonstrate adaptability; reveal unexpected behaviors.                       |
| Legal, Compliance *Harder*  | Process walk-throughs, actual documentation, real reviews Live process demonstrations, actual case walk-throughs        | Use anonymized real scenarios; show time requirements; demonstrate decision frameworks.              |
| Finance *Harder*            | Working financial tools, actual analyses, decision frameworks Live data analysis, real-time forecast demonstrations     | Use real company data; show edge cases; demonstrate how decisions would actually be made.            |
| HR, People *Harder*         | Process experiences, communication materials, actual tools Role-played experiences, real tool interactions              | Demonstrate the emotional journey; show actual timeframes; use realistic scenarios.                  |