Introduction

There’s a moment every insights leader recognizes.

The room is excited. The demo is dazzling. The roadmap is already being mentally celebrated. And somewhere in the back of your brain, a quieter voice is saying, “We are about to build the wrong thing… just faster this time.”

So it was just around 14 years ago that I had that same feeling in a very different setting.

I was on an unexpected social visit to meet Gordon Moore, retired cofounder of Intel, philanthropist, and, charmingly… humorist, at his home in Kona, Hawaii. Late afternoon. The kind of light that makes you question every “important” meeting you have ever attended in a windowless room.

You know the kind of view I mean. The ocean looks infinite. The sky looks like it has a marketing budget. And your brain starts doing that dangerous math like, “If I moved here, could I make a living… appreciating sunsets professionally?”

Approaching the home via a winding driveway through an ancient black lava field and walking up to the door over a number of highly imagined stepping pads past a glorious coy pond, I felt a bit anxious and thrilled at the same time at my unusual fortune to have the opportunity to even be there in this moment. Surreal to the core.

Upon greeting us at the door, Gordon and his wife, Betty, were instantly charming, warm, and gracious to welcome us for this visit, and that late August 2011 afternoon, I had the chance to sit and talk with Gordon Moore, co-founder of Intel and the mind behind what the world later called Moore’s Law.

This particular meeting was about three friends (my host and the Moores), discussing Gordon and Betty’s annual donation to a local community hospital. I was taken by their generosity and genuine care for their community. Something I will never forget.

His home was beautiful, yes. Kona is unapologetically beautiful.

But what stuck with me was not the view.

It was the way he talked about technology and markets like they were one integrated system, not two departments that occasionally exchange emails. It was at once clear why giving back was also wired into his philosophy.

Gordon was relaxed, understated, and quietly funny. At one point, he looked out at the water and said something like, “Business travel used to be Philadelphia in February. This is… harder to complain about.”

That is Gordon in one sentence. A gentle truth, delivered with a little grin, and no need to make it dramatic.

Of course, I asked him the question everyone asks.

“Did you realize you were creating Moore’s Law and what that meant?”

He smiled like he’d heard it a thousand times, but in a generous way. Like, he still found it amusing that people treat a chart like scripture.

“I wasn’t trying to write a law,” he said. “I was describing a trend I could see. The bigger point was what happens when the economics keep working.”

Then he said something that landed harder than the famous line ever could.

“The number is interesting. The insight is what matters.”

That sentence has followed me ever since, especially now, living inside the generative AI moment where numbers are everywhere, and certainty is… optional.

See below: The legend Stan Sthanunathan, breaking down the nature of “What is Insight”.

So I asked Gordon what he meant by “insight.” Because we all say it. We all sell it. Half the time, we put it in a box and ship it as a deck.

He leaned forward slightly and said, “In R&D, physics tells you what’s possible. Insights tell you what’s worth doing.”

And then he gave me the most practical framework I’ve ever heard, delivered like he was describing how to make coffee.

“You’re always answering three questions.

Can we build it?

Can we build it reliably?

Can we build it so someone will buy it without crying?”

I laughed because it was funny.

Then I realized it was also brutally accurate.

The “crying” is not literal. It’s what happens when the real world meets your product, and someone pays the price for tradeoffs you did not face early enough.

It’s the customer crying because it’s too expensive, too complex, too risky, too hard to integrate, too unpredictable at scale.

It’s the internal team crying because the roadmap was built on assumptions that did not survive contact with reality.

It’s the business crying because adoption is not a feature. It’s an outcome.

Sometimes the crying is unavoidable.

But it is almost always predictable.

And in a compounding world, that is the job of insights. Not reporting. Not decorating. Navigation.

Here’s the part we tend to forget when we get dazzled by a new technology curve.

Moore’s Law did not change the world because chips got faster.

It changed the world because constraints fell.

When computation got cheap enough, small enough, reliable enough, and accessible enough, whole categories of behavior became normal. The magic was never the silicon alone. The magic was what the silicon removed.

I used to think Moore’s Law was about speed.

Then Gordon gave me that line about crying.

And I realized the real story was adoption.

Technology scales when capability compounds.

Adoption scales when constraints disappear.

That’s the AHA. That’s the whole thing. Put it on the wall. Tattoo it on a roadmap. Stitch it on a pillow if you have the kind of house where that makes sense.

And now we are living through the next version of that story.

Generative AI is having its Moore-like moment. Capability is improving quickly. The demos are getting more magical. Every week there’s a new model, a new breakthrough, a new “this changes everything” headline.

But if you are building products, running an insights team, or steering a business, you can feel the gap.

Capability is not the same thing as adoption.

In fact, the fastest way to get confused in 2026 is to mistake a better model for a better outcome. It’s benchmark theater. It looks like progress. It feels like motion. It’s also a great way to ship something nobody uses.

Here is what the “crying” looks like in the generative AI era.

The CEO is excited.

The demo works.

Someone says, “This is going to transform everything.”

Then the compliance person goes quiet, which is never a relaxing sign.

The head of CX asks, “Who owns output quality?”

Support asks, “What do we tell customers when it confidently makes something up?”

IT asks, “Where does the data go?”

And suddenly, everyone realizes the truth.

The model is impressive.

The workflow is not ready.

That’s the constraint.

And customers do not adopt generative AI because it has gotten smarter on a benchmark.

They adopt it when something becomes frictionless enough that they can use it without holding their breath.

They adopt it when it becomes safe enough that it does not feel like gambling with their job.

They adopt it when it fits into a workflow instead of becoming another tab, another tool, another “please update your password” moment.

They adopt it when the value shows up in minutes, not in a six-week pilot that ends with a nice PowerPoint and no behavior change.

This is where I keep hearing Gordon’s voice in my head, in that calm Kona cadence.

Physics tells you what’s possible.

Insights tells you what’s worth doing.

And in the generative AI era, “worth doing” is not the coolest feature you can build.

It’s the constraint you can remove so completely that the customer says, almost casually, “Oh. Of course, I’m using that now.”

That is the inflection point. Not faster. Not bigger. Not fancier.

Just easier.

Safer.

More predictable.

More adoptable.

When the curve is steep, there are a hundred possible things you could build. Most will be impressive. Some will be useful. A few will become inevitable.

Our job is to find the difference before we spend the next twelve months building ourselves into an expensive hobby.

Here’s my private rule for teams building in a compounding moment.

If we can’t name the constraint we’re removing, we’re not building a product.

We’re building a demo.

So what does an insights leader do right now?

Not write longer reports.

Not create prettier dashboards.

And certainly not host more share-outs where everyone nods politely and then goes back to shipping what they already planned to ship.

We make the “why” unavoidable.

We put the customer’s constraints in the middle of the conversation, where nobody can ignore them.

We translate reality into decisions that hold up when the market starts throwing chairs.

And we keep asking the question Gordon Moore, knowingly or not, handed me that afternoon in Kona.

Not “Can we build it?”

We can build almost anything now.

The real question is: what are we building that removes a real constraint so completely that adoption becomes natural?

Because that’s how Moore’s Law became a force of history. Not by increasing speed, but by shrinking friction.

And that’s how generative AI will transform businesses, too. Not when it gets cleverer, but when it becomes safer, simpler, and embedded in how people already work and live.

Kona was infinite.

But exponentials are not.

If this curve keeps scaling for the next twenty years, a lot will become cheap, instant, and expected.

The only question that matters is whether we’re building toward that future, or polishing value that’s about to become invisible. Chatbots and Conversational A/I are not nice-to-haves; they are rapidly becoming invisible because they are need-to-haves. The real question is where are customers, organizations, and individuals struggling today with everyday pain points, and what can we do to alleviate them? We can build it once we understand this. This has never been the issue.

That is Discovering Why.

Not discovering what’s possible.

Discovering what will be adopted, what will endure, and what will matter before the crying starts.