The Smartest People in the Room Have Already Been Wrong in Front of Everyone

Why “learning in public” can seem like a weakness, but is actually the only credential that has lasted through five tech cycles

You know the drill. You’re supposed to act like you’ve got it all figured out. You cram the night before, walk into the meeting with your bullet points, and pretend you didn’t just Google half the jargon on the way over.

That used to work. Back when the world moved at the speed of dial-up, you could disappear for a few years, come back, and still be The Expert. As if expertise was a place you could just move into and never have to leave.

That deal is off. And not in a philosophical way — in a practical, this-affects-your-Monday way. The tool you mastered last quarter just shipped a feature that made half your workflow obsolete. The only people who caught it were the ones openly fumbling with it, while you were busy looking competent.

So I want to talk you into something that may terrify you: let people see you when you don’t know the answer. Learn in public. Build in public. Let them watch you in the middle of figuring it out, judgment and all.

(I’m about to do exactly that — my own Learn in Public notes are in Substack Notes. Consider this the argument for why you should do it too.)

Why do I get to tell you this? Because I’ve been wrong in public for twenty-five years.

The Then

In 1999 I was running a Usenet group like it was a small nation-state and posting my entire interior life to a personal blog under my real name. I did not know what I was doing. Nobody did — the medium was four years old. We were all just typing into the void and finding out, together, in front of each other, which things were embarrassing only after we’d already published them. Did I mention it’s really difficult if not impossible to remove old Usenet forum listings?

Then the dot-com boom, where I had opinions about the future of the web that read, in hindsight, as adorable. By Web 2.0 I was confidently right about some things and confidently wrong about others — and both are still searchable.

The one I’d point to is OpenSocial. In 2007, Google rallied basically everyone who wasn’t Facebook — MySpace, Orkut, LinkedIn, a dozen others — around a shared standard so you could build a social app once and run it everywhere. I was sure this was the future. Open beats closed! A win for open standards beats a company’s walled garden! It’s the bet that should win, the one that sounds smart at every meetup, right? Wrong. Facebook’s single closed network devoured all of them. Most of those host sites don’t exist anymore. The standard got quietly handed off to a foundation and went to die.

The cost of being wrong in public, which everyone fears, was real every time. People noticed. Some remembered. A few even mentioned it later.

But it didn’t matter. What mattered was that the people who made mistakes openly were the ones who truly learned. Those who stayed quiet to protect their image, waiting until they felt sure, faced every new change as if they were beginners. But now, they also had to pretend they weren’t beginners.

It is no longer possible to quietly become an expert in rapidly evolving fields. Expertise now resides in the learning process itself, which must be visible to be effective and cumulative.

The Now

Here’s the trap specifically for someone in your position.

You’re at the director level or higher. You’ve earned authority by seeming to have all the answers. Now, you’re asked to make decisions about AI tools quickly, over and over, often without enough context. In this environment, admitting you don’t fully understand something can feel like admitting you shouldn’t be making a decision.

So your instinct is to do what has always worked: evaluate in private, present your findings with confidence, and never let anyone see you reading the documentation. You want to protect your authority.

The problem is that this approach is now risky rather than safe. If you only share your conclusions, three things happen:

1. Your team can’t see your reasoning, so they can’t catch your mistakes before they become policy.

2. You miss out on feedback from others, like the colleague who says, “oh, that tool can’t actually do that” before you commit budget to it.

3. And you show everyone that not knowing is something to hide, which means they’ll keep their own confusion secret until it causes problems.

Working in public, by sharing your half-formed evaluations, your “here’s what I tried and here’s where it broke,” and your honest “I’m not sure yet,” does the opposite. It turns your mistakes into shared knowledge. It invites others to help catch your errors early. And it gives your team permission to be openly confused too, which is the only way an organization can learn quickly enough to keep up.

What feels like exposing weakness is actually what builds your organization’s collective intelligence. Sharing only polished conclusions may feel safe, but it quietly leaves everyone less informed, including you.

The Calibration

Here’s what you can actually do on Monday morning, especially if you can’t just blog your whole AI evaluation process for everyone to see and call it a strategy.

Start small. You don’t need to learn in public for the whole internet. Just do it with your team, a smaller, more impactful group. The next time you evaluate a tool, share your process in a channel your team can see. Say something like, “Trying X this week. Here’s what I’m testing it against. Will report back.” Then follow up, including the parts where the tool let you down.

Separate the two fears, because they’re not the same. One fear is “I’ll look incompetent.” The other is “I’ll be held to something I said while I was still figuring it out.” The first one is mostly in your head — people respect visible process more than you think, and the leaders who never show it read as either fake or out of touch. The second is a real, manageable risk: you handle it by labeling the work as in progress. “Current thinking, subject to change” is a complete sentence and a full defense.

Distinguish learning in public from performing in public. This matters, and most LinkedIn advice collapses it. Learning in public is showing genuine in-progress work so it can be corrected and shared. Performing in public is manufacturing fake vulnerability — the “here’s my humbling failure (that actually makes me look great)” post — to harvest engagement. Your team can smell the difference instantly, and the performed version poisons the well for the real thing. The test: are you sharing this because it’s useful to someone watching, or because it’s flattering to you? If it’s flattering, keep it in the drafts.

See your past mistakes as valuable experience. If you’ve been through any previous tech cycle—and as a director, you have—you’ve already learned in public, whether you meant to or not. Maybe it was a CRM migration that went wrong, or a “digital transformation” that didn’t change anything. You got through those by making mistakes where others could see and then adjusting. You already know how to do this. You just stopped seeing it as a strength.

The reason I can write a newsletter about being early to the future yet still be seen as late is that my record of mistakes is the whole point. Anyone can claim foresight after the fact. The only credential that has lasted is the visible, searchable, and sometimes embarrassing trail of someone figuring things out in real time and being willing to be wrong when it matters.

The future is, once again, happening slightly faster than any of us can become experts in it. The people who’ll do best aren’t the ones who figure it out first. They’re the ones who let everyone watch them as they don’t have it figured out yet.

That’s not a consolation prize for showing up “late-ish” — they’re the ones winning the whole damn thing.