The ROI on Trusting Your Own Thinking

How AI creates space for thinking without consequence

Your best thinking is sitting in thoughts or drafts you're not sure are ready.

You know the ones. Half-baked ideas you've been tinkling with for weeks. Notes you wrote at 11pm and haven't looked at since. Scribbles you can’t decipher. Questions you'd ask if you were certain they weren't stupid.

They're waiting for something. Permission, maybe. Or proof that the thinking is worth taking seriously.

Proof is seen daily; I've been watching it unfold on Substack.

People show up tentative. They post a Note, half-question and half-statement, testing whether the thing they've been mulling actually holds weight outside their own head. The writing is careful. A little hedged. You can feel them bracing for the silence, or worse, the correction.

Then something surprising jogs your Substrack scroll.

A few people respond. With recognition. I've been thinking about this too. Or this landed for me. Just evidence that the thinking registered somewhere beyond the screen.

And then you watch them write again. A little less hedged this time. A little more willingness to follow the thread before they know where it goes.

Over weeks and months, the tentativeness burns off. The voice sharpens as they begin to trust their own thinking. One small confirmation at a time.

I've felt it myself on this platform in so many tiny, little ways. And I've started to understand what's happening beneath the surface: The Substack space matters more than it should…

Substack, for whatever else it is, offers something quietly radical: room to think out loud without consequence. Enough distance from the immediate social cost of being wrong that you can actually risk being interesting.

That's a big thing. Most environments punish exploration in real time. You float an idea in a meeting and watch people's faces. You share a half-formed thought with a colleague and feel them filing it away, updating their model of how seriously to take you. The stakes are invisible, but they're real.

So most people stop. Gradually. They learn which thoughts are safe to voice and which ones carry too much risk. The range narrows. The thinking gets smaller. And after enough years of that, you forget you ever had more to say.

What I've been watching on Substack is people remembering.

Christina Pearl wrote something recently that stopped me mid-scroll.

She pointed out that AI quietly gives back confidence. The courage to try things you'd normally postpone. To ask the questions you've been sitting on. To explore unfamiliar territory without announcing it publicly. To make connections between seemingly unconnected topics.

If we only measure AI by productivity or cost savings, we miss that entirely.

Emma Klint built on it. AI as practice ground, she called it. A space that meets you where you are and allows you to express yourself the way you need. Where the questions and dead-end tangents are just part of the process…

Something happens when you give yourself permission to be yourself. You start trusting your own thinking.

Christina & Emma, and the many others who joined the chat, were pointing at the same thing I've been watching unfold in public. The same thing that happens in private when someone opens a conversation with an AI and realizes nobody is keeping score. Here’s what the productivity metrics miss entirely…

When we talk about AI ROI, we talk about time saved. Tasks automated. Output increased. All real. All measurable. All beside the point. The Point: The deeper return is a person who now believes their thinking is worth doing. That belief changes what someone attempts. It changes what they say in the meeting, what questions they ask, and what problems they're willing to wrestle with instead of route around. A person who trusts their own thinking shows up differently than someone still waiting for permission.

That difference compounds. It shows up six months later when they propose something nobody else would have proposed, because they'd been practicing the thinking all along. In spaces where no one was watching. In conversations with an AI that never once made them feel stupid for asking.

There's a version of the AI conversation that lives and thrives in the weeds.

• Write this way...

• Scheduling concerns...

• Which tool is best...

• Who's the guru this week...

That conversation has its place. But, there's another version that many people never reach because the weeds are so thick.

It's the version where you lift your head and realize that the thing AI does best is giving you a place to discover you had better questions than you thought. A place to think without the weight of being watched. A sandbox where the thinking you do is real even when the stakes aren't.

The writers I've watched find their voice on Substack found it by having a space that let them hear themselves. One post at a time. One Note at a time. One quiet confirmation that their thinking mattered to someone, even if that someone was just them.

AI offers the same thing, only private. Only yours. A room where you can think out loud until you're ready to think out loud in rooms that cost you something.

So here's the invitation…

Forget the prompts for a minute. Forget the debates about which model is smarter. Forget the gurus and the certainty merchants.

Ask yourself when you last trusted your own thinking enough to follow it somewhere you couldn't see.

If the answer is "a while ago," you're just out of practice. And practice requires a place where being wrong costs you nothing except the chance to be less wrong next time.

That place exists now. Several of them, actually.

The only question is whether you'll use them to get faster at output or to get better at thinking.

One of those has an ROI you can measure.

The other has an ROI that changes what you can measure in the first place.