The Recursive Loop: How Sharing Creates What's Worth Sharing

Most people think you share after you've made something worth sharing.

The truth? Sharing is how you make something worth sharing.

That reversal—the idea that exposure precedes excellence—might be the single biggest unlock for creativity in the AI era. It's certainly changed my life.

The Dinner I Didn't Want to Attend

Last night I was tired. Run down. Ready for bed.

But the calendar said otherwise, so I showed up to a dinner I really didn't want to attend.

Two hours later, I'd learned fascinating details about Amazon's internal innovation mechanisms I'd never encountered in any book or article. The kind of insider knowledge you only get from someone who was actually there, doing the work.

Here's the thing: I didn't get lucky with my dinner companion. I got curious because I have a blog.

That might sound strange, but it's true. The commitment to write weekly has fundamentally changed how I experience the world. I wasn't just enduring a dinner—I was unconsciously scanning for raw material. By dessert, the obligation had become treasure hunting.

To a writer, everything is grist for the mill. And that's not just a poetic phrase—it's a perceptual shift. When you've committed to producing output, suddenly everyone you meet becomes potentially interesting, because you need material.

The dinner didn't become interesting because it was interesting. It became interesting because I'd trained myself to look for oysters.

The Loop You Didn't Know You're In

Creators have understood this dynamic for generations.

Lorne Michaels built it into Saturday Night Live's DNA: "The show doesn't go on because it's ready—it goes on because it's 11:30 p.m. on Saturday."

Thomas Edison set invention quotas. Jerry Seinfeld filled yellow legal pads every morning. Victor Hugo stopped mid-sentence at dinner to jot down thoughts in his notebook.

All of them built systems that forced iteration before perfection—what I call the Recursive Loop:

The commitment to share creates the necessity to experiment.
The experiment creates material worth sharing.
Sharing reveals what to try next.

It's not "do good work, then show it."

It's "show your work, so you'll do good work."

This dynamic is as old as Hugo's notebooks and as new as AI model training. The same logic that improves GPT improves humans: exposure → feedback → refinement → better next round.

Your audience is your dataset.

Why This Matters for AI Adoption

Every leader asks: "How do I get my team to actually use AI?"

They imagine they need more training, better tools, clearer guidelines. But the missing element is rarely knowledge. It's commitment architecture—a reason to try, share, and reflect in public.

Take the AI Junto, our local community of practice.

Each week, members post experiments they've run. Not necessarily successful experiments—just what they tried, what they learned, what questions emerged.

And here's what happens: seeing others’ experiments creates that gentle social pressure—I should try something. I don't want to be the person who never shares.

That mild discomfort? That's the engine.

It's not guilt—it’s accountability. The commitment to share doesn’t just document experiments; it creates the necessity to experiment.

And the experiments done because of that commitment often become the best ones. Thinking about what will be useful to others makes you more thoughtful about the work itself.

The commitment to demonstrate creates better demonstrations.

Organizations usually imagine this sequence:
Train people → People experiment → People share results (maybe)

But reality works like this:

Create commitment to share → People must experiment → Experiments improve → Sharing becomes easier → Commitment strengthens

You don't inspire experimentation and then hope people share. You build social architecture where not sharing feels uncomfortable.

Leaders need to demonstrate their AI use (read the Start with the Executive Team paragraph in Make AI Adoption Non-Optional). Not just to model behavior, but because the act of demonstration forces them to experiment, which yields better use cases, which reinforces the habit.

It's a recursive loop. The output creates the input.

This is why conveners like Cheryl Eckhardt are so powerful in organizations—not because they're technical experts, but because they create persistent social environments where people commit to sharing. That commitment manufactures the experimentation that produces the breakthroughs.

One person’s experiment becomes proof for others. The loop scales from personal to collective—until your whole team behaves like a learning organism, training itself on its own shared experience.

The Courage to Go Premature

But here's the hard part: nobody ever feels ready to share.

I see this constantly in design workshops, especially among working professionals. I'll ask, “Who feels ready to share their work?”

Almost always, not a single hand gets raised.

Why? Because perfectionism runs deep in our veins. We want our work to be perfect before unveiling it to the world.

Then we share anyway.

After the feedback and critique, I ask: “Whose work improved because they shared?”

Every hand goes up.

So no one was ready, and everyone benefited. What gives?

In most work, we've been conditioned to make it perfect before sharing. But with creative, exploratory work, the opposite is true: share it in order to make it more perfect.

As James Webb Young wrote in has fantastic handbook A Technique for Producing Ideas nearly 100 years ago:

"Do not make the mistake of holding your idea close to your chest at this stage. Submit it to the criticism of the judicious. When you do, a surprising thing will happen. You will find that a good idea has, as it were, self-expanding qualities. It stimulates those who see it and add to it. Thus possibilities in it which you have overlooked will come to light."

Reid Hoffman reframed it for startups: "If you're not embarrassed by the first version of your product, you've launched too late."

The same applies to AI experiments. Launch early. Share imperfectly. Let the loop do its work.

The Share-to-Learn Loop

Here's how the recursive loop actually works in practice:

1. Commit to Output
Choose a frequency that scares you a little—weekly, biweekly, monthly. Schedule it now.

2. Expose the Work
Ship rough, not ready. Ship on schedule, not on perfection. The commitment is to the rhythm, not the result.

3. Invite Response
Treat feedback as free R&D.

4. Reflect & Refine
Each share becomes new training data—for you, for your team, for your AI collaborator.

5. Repeat Relentlessly
Consistency beats brilliance. Loops beat leaps.

This is how AI models improve. And it's how humans improve through community.

Your Recursive Loop: Three Ways to Start

Option 1: Start a Weekly AI Share-Out

Text three peers right now: "Want to start a weekly AI Show-and-Tell? 5 minutes each, one experiment per week. No successes required—just what we tried."

Schedule the first one. Post a takeaway publicly afterward.

Option 2: Join an Existing Community

The AI Junto has weekly share-outs and a persistent channel where people post experiments between meetings. You don't need to build the architecture—just step into the loop.

Or start a simple Slack thread or “AI Office Hours.” The key is recurring commitment, not one-off events.

Option 3: Lower Your Bar for "Worth Sharing"

The biggest barrier isn't capability—it's perfectionism. You think, "This experiment didn't work perfectly, so I shouldn't share it."

Wrong.

Failed experiments are often more valuable to share than successful ones, because they help others avoid the same pitfalls. The Junto's most engaged discussions happen around experiments that half-worked or created unexpected problems.

Share the messy middle. That’s where breakthroughs hide.

What Changes When You Measure Loops, Not Participation

If you lead a team, stop asking "How many people tried AI this week?"

Start asking:

• How many people shared something this week?

• How many experiments were inspired by those shares?

• How many new questions emerged?

These are the metrics of a learning organization—an organization that, like a good AI model, keeps training itself on its own collective experience.

You're not measuring individual adoption. You're measuring whether the loop is running.

Because in five years, organizations won't be distinguished by who had the most advanced AI tools. They'll be distinguished by who created the most effective communities of practice around them.

And communities of practice are just recursive loops at scale.

From Obligation to Opportunity

When you’ve built output commitments into your life—a blog, a weekly share-out, a monthly demo—you stop enduring obligations and start hunting for material.

The boring dinner becomes a conversation safari.
The failed experiment becomes a teaching moment.
The random insight becomes tomorrow’s post.

You’re not waiting for inspiration.
You’re engineering serendipity through commitment architecture.

Because here’s what I’ve learned—from a year of daily blogging, from the AI Junto, from helping organizations transform:

Don’t wait until you have something worth showing.
Showing is how you make something worth having.

Exposure precedes excellence.
Consistency beats brilliance.
Loops beat leaps.
Stop waiting for perfect.
Start sharing.
Let the recursive loop do its work.

Related:
Set an Output Schedule
Capture Inspiration
Punish Inaction
Sharing Before You Feel Ready
The Most Important AI Role Has Nothing to do with Code
Practice in Community
Best $5 Book on Amazon: A Technique for Producing Ideas

Want to build your AI practice in community? Join the AI Junto—where commitment creates capability, and sharing creates what's worth sharing.