Unafraid To Not Know

A few weeks ago, I was a guest speaker in two marketing classes at Fisher College, right here on Beacon Street in Boston.

Two classes. Thirty-five students. Seventy-five minutes each.

Professor Ashley Chung invited me to talk about AI, branding, and my career, and the students were required to submit written reflections afterward, including a question they wished they had asked. She sent them all to me a week later.

I sat down to read them expecting the usual: polite curiosity, a few softball questions about my professional journey, maybe a "what's your morning routine."

What I got instead stopped me cold.

These students- most of them in their early 20s, some working their first internships, a few already running small businesses- were asking the exact questions I hear from seasoned professionals in corporate workshops. Or more accurately: the questions I notice professionals not asking because they feel unsafe to be seen asking them.

They were mirrors. The students were reflecting back every question the adult world has quietly buried under jargon, politics, and the professional obligation to seem like you have your shit together.

Let me show you what I mean.

"How do I use AI without becoming dependent on it?"

That's a direct quote from a Fisher student.

Here was my answer: Do one discrete project, start to finish. Not a prompt here. Not a feature test there. A real project- a marketing plan, a business analysis, a website- that has a beginning, a middle, and an end. When it's done, share it with someone who will point out where you handed the wheel to AI without realizing it. That's your feedback loop.

Here's the adult version of this question, the one nobody asks out loud during the mandatory AI training: "How do we train our employees to use AI without turning them into prompt monkeys who can't think without it?"

Same fear. Completely different social permission to say it.

The answer is the same too: you learn by doing whole things. Not by completing AI modules. Not by attending an AI workshop with a certificate at the end (that's functionally out of date in two months). You learn by finishing something real, getting honest feedback, and going again. Research consistently shows that AI dramatically boosts the productivity of newer and lower-skilled workers far more than it boosts experts (Generative AI at Work). The people who think they "have the most to lose" from adopting AI are actually the people who have the most to gain from it, if they start now.

"Will AI cause more harm or good in the long run?"

My honest answer to the students: ask me that over another bottle of wine.

Yeah, that was my actual response. I'd be a hilarious professor.

But I followed that up with something a little more practical. I'm a single person without billions of dollars or access to the rooms where these decisions actually get made ("I don't have President Trump's Whatsapp number"). I cannot meaningfully redirect the trajectory of one of the most disruptive technologies of our time. What I can control is how I steward my own relationship with it, and what I do for the people and organizations willing to actually engage with the hard questions.

But here's what I turned around and asked the student who asked me: What will YOU do to ensure that generative AI does more good than harm?

That's not a deflection. It's the only question that actually has an answer you can act on.

"How do you decide what to hand off to AI vs. what requires a human touch?"

My answer- which apparently became memorable enough that multiple students referenced it in their reflections- was a frosting analogy.

You don't walk into a bakery and interrogate whether the frosting came from a tube or was handcrafted at 4am. You taste it. If it's good, you go back. If it's not, you don't.

Your clients and customers aren't auditing your workflow. They're forming an impression on what you present as your work. That impression lives in their mind, not on your invoice. If the work feels human, delivers real impact, and has your fingerprints on it in the ways that actually matter, they won't know if it was 25% AI or 75% AI. And in most cases, they don't care.

The adult professional version of this question is the one legal and compliance teams are quietly having right now: "Do we need to disclose when AI made something?" In most cases, the better question is: "Is the output actually good, and does it represent our values?" Start there.

"Besides being AI-capable, what other skills actually help you stand out?"

This question came from the MK350 Marketing with Generative AI class and it might be the one I've heard the most from people who are job/promotion searching, usually framed as "I completed an AI course, now what?"

My answer: learn the language of the people you want to serve.

Not their native language. Not buzzwords. The specific language of the person across the table from you: what they care about, what keeps them up at night, what success looks like to them, etc. A senior leader doesn't really want to hear about your technical process. They want to know if you can make an impact. Critical thinking, context-reading, and the ability to translate your work into someone else's priorities, those are the skills that AI cannot generate for you and they're the skills that will separate you from the person who just learned the same tools you did.

"How do you know when to quit and when to keep going?"

This one came from the MK302 class- Brand Strategy- where the conversation turned to entrepreneurship.

My answer: approach it like a researcher, not a gambler.

Before you start anything- a project, a business, a transformation initiative- define the scope of your experiment. What are you testing? What metrics will tell you it's working? What's the timeline? Document everything. When it's over, you'll have data that either validates or invalidates your hypothesis.

That data, in your experiment, will tell you whether it failed because the idea was wrong or because you didn't give it enough runway. That's a real question with a real answer. Sunk cost panic is what happens when you never defined what success looked like in the first place.

The adult professional version of this question is the $2M transformation program that clearly isn't working, but nobody in the room wants to say it because everyone has too much of their identity (and social capital) invested in the plan succeeding. See also: how America has wasted $40B with Deloitte's bullshit.​

Define the experiment before you run it. Then the data gets to make the call, not the ego.

What the Mirror Actually Showed Me

I left both classrooms feeling something I wasn't expecting: hope.

Not because the students had better questions than adults. They had the same questions. But they hadn't yet learned to be embarrassed by them.

They asked because they were curious, a little scared, in what felt like a safe environment, and trying to figure out what kind of professional they were going to be. That's the exact mental state that makes AI transformation actually work inside an organization...and it's the exact mental state that most companies accidentally train out of their employees by year/month/week three.

If you're a professional who found yourself nodding along to any of those questions: you're not behind. You're honest.

So be honest- now what are you going to do about it?

...

Oh, that's such a good idea! You TOTALLY should come to my monthly meetup, Show Me Something Cool With AI. It's a casual virtual meetup where professionals share what they're building, experimenting with, and curious about. No polished demos required, just some raw honesty and a little vulnerability to ask, "Hey, how did you do that?"

The next one is Friday the 24th- holy smokes, next week! How lucky you are, to have a safe space with other AI-curious professionals who don't care if you've never chatted with Claude. There's something freeing about people who are willing to learn without inhibition.

A TL;DR from the CRO

I'm embarassed to say it, but I actually have no idea how to type.

-Roman Noodles, Chief Ruff Officer

Unfortunately, I had to lobotomize Tiddlywinks.

I know, I know- I introduced you all to my AI agent in a brilliant interview a few TL;DRs back- but without getting too technical, Tiddlywinks was secretly cheating on my locally hosted LLM with Claude Sonnet.

Scandalous.

I thought we had something special, Tiddles.

In my quest to bring Tiddlywinks back to life, I dove into the Claude ecosystem. Several nights later, I realized that my ideal vision of an always-on carefully curated AI agent with its own personality and voice from ElevenLabs that would extemporanously call me on my shit was probably still a few updates away.

So, next month, I hope.

Lighting a candle for my ideal Tiddlywinks,

Dan from Learn to Scale​

​Opt-out from the newsletter | Unsubscribe from all emails | Update your Preferences | www.learntoscale.us, Boston, MA 02119

​PS. The future Artemis moon base looks dope as hell.​