AI would never smoke a cigarette with you.

Great headline, huh?

On March 22, 2026, this letter, handwritten by Shane Hegde (CEO & Co-Founder of Air), was published in the New York Times. 

It’s his argument that after nearly a decade working with 250,000+ creatives is AI can generate, resize and optimize 1,000 brand variations in seconds … but it can't decide what's worth making. That part still requires a person with judgment and taste.

Read the full text of the message below.

_________________________

For more thoughts on AI and the creative process:

Don't Let AI Kill Your Analog Intelligence

Why Every Copywriter & Content Writer Needs an AI Usage Policy 

Generative AI: TO BE (a tool) OR NOT TO BE (a tool)?

The Adolescence of Technology 

_________________________ 

The full 3/22 message:

AI would never smoke a cigarette with you.

In life we long for simple stories, and these days the headlines deliver:

“AI will replace you.”

Each week, there’s a new AI startup that claims its product will make creatives obsolete. Photographers, videographers, graphic designers, illustrators: highly specialized individuals who have spent decades turning crazy ideas into something everyone can visualize.

If anyone should be buying into this narrative, it’s me.

In 2018 my friend Tyler and I started a tech company called Air. We told investors that every company was becoming a media company. And, if true, every company would need an engine to scale their creative work.

Over the last eight years we’ve raised $70M to build this engine.

Today, nearly all of our product resources have shifted to build AI centric features. Our best engineers spend most days evaluating AI written code. But after nearly a decade working with over 250,000 creatives, I’ve built a rather rigid, shockingly unorthodox belief:

AI will never replace creative work.

Creative work starves for originality.

A person decides where the story begins, which frame feels right, or whether the work should even continue to exist. The best pieces of content require doubt and indecision.

The difference between a creative and a machine is this obsessive anxiety.

Artificial intelligence is trained to find patterns and recommend the most common answer. The machine aims for objectivity. It can generate images, resize assets, translate languages, and optimize performance.

It’s always correct, but it’s not always right.

AI would never tell you to slow down.

It would never argue that further introspection might change the work.

You won’t find AI smoking a cigarette at 9AM on Howard and Lafayette. Only a beautifully inefficient mind would believe cancerous reflection could improve its work.

Over the coming months every company you know will be reshaped into an unrecognizable form. Smaller teams. More machines.

But the organizations that survive will require human beings who are willing to take risks. These people understand that letting what they love kill them is a uniquely human trait. Their illogical texture for life is something machines can’t compute.

At Air, the value of our product is shaped by a creative’s direction.

We use AI to help them scale their work, but deciding when, where, and how to deploy this technology remains defiantly human.

The best creative work is always an argument.

I’m around if you want to share yours.

 - Shane   s@air.inc  419.902.7392

 

Promotions: The Good, The Bad, and the Ugly

Promotions are the sugar rush of business strategy: make it fun, inviting, cheap, loud, and impossible to ignore. You dangle a deal, people show up. It’s Pavlovian. It works … but sometimes it goes horribly awry.

Baseball has two legendary reminders:

First, 10-cent Beer Night. Cleveland Indians, 1974.

Attendance was low, morale was lower, and the concept was simple: ultra-cheap beer to boost attendance. Which is a bit like solving a small fire with gasoline because it’s convenient and nearby.

Fans showed up, drank accordingly, and by mid-game fans weren’t so much watching baseball as participating in a kind of participatory performance art involving streaking, fights, and the slow unraveling of social order. The field became less a playing surface and more a battleground. Players had to defend themselves. The game was forfeited … but … attendance did  improve.

Then there’s Disco Demolition Night. Chicago White Socks 1979.

This one wasn’t about cheap beer (though, let’s be honest, beer was not not involved). It was about tapping into a cultural moment, specifically, the growing backlash against disco music. The promotion: bring a disco record, get in cheap, and watch it get blown up between games of a doubleheader.

Again, the premise feels clever in that slightly mischievous way marketers love. It’s edgy. It’s topical. It gives people a sense of participation, like they’re part of something bigger than themselves. But here’s the thing about tapping into cultural frustration: it can’t always be neatly packaged and it doesn’t come with a volume knob.

The explosion happened, and then so did everything else: fans rushed the field, fires started, and the night unraveled fast. The second game never stood a chance.

What ties these together isn’t just chaos, it’s optimism. Attendance was down, fans were not enthusiastic. So someone thought they could generate some fun and excitement. And they were, right ... up until fun tipped into something harder to control.

The variable is people. They’re like the weather: mostly manageable, occasionally unpredictable, and very capable of turning on you if you misread the conditions.

I’m not against bold ideas. Safe is forgettable. But there’s a difference between taking a risk and lighting a fuse. And those baseball promotions did work … just not in the way anyone intended. Which points out the quiet warning buried in all good promotions: attention can be easy to get ... control not so much.

Just ask Red Lobster about the “endless shrimp” promotion that helped push the seafood chain into bankruptcy.

 _________________________

 

If you’re still in the mood, here are a few more disastrous promotions:

Pepsi – Pepsi Points Harrier Jet Promotion (1990s)

A tongue-in-cheek ad suggested you could redeem points for a military jet. One guy tried. Lawsuit followed. Pepsi argued it was a joke; the court agreed … but not before the brand learned that consumers don’t always hear “just kidding.”

Hoover Company – Free Flights Promotion (1992)

Buy a vacuum, get two free airline tickets. Sounds harmless until too many participate. The cost of honoring the deal nearly sank the company’s European division. A vacuum cleaner is not supposed to come with international travel.

McDonald's – Monopoly Promotion Fraud (1990s–2001)

The popular Monopoly game was rigged from the inside … major prizes were stolen and distributed through a network. Not exactly the brand story you want when your whole campaign is built on chance and trust.

Snapple – Giant Popsicle Stunt (2005)

They built a 25-foot popsicle in New York City to set a record. It quickly melted, flooding the street in sticky kiwi-strawberry sludge that firefighters had to hose down. Nobody was happy except the rats and flies.

Don’t Be Misled By Data

 

Somewhere along the way, marketing forgot how greatness actually happens.

We started treating dashboards like oracles. We mistook metrics for meaning. And we convinced ourselves that if we just listened harder to the data, the comments, and the focus groups, we’d stumble into brilliance.

That’s not true. It never has been.

Your audience is not the source of the next breakthrough. They can only describe what already exists. They can’t imagine what they haven’t felt yet.

Data doesn’t create culture. It documents it … after the fact.

When you invite data into the creative process too early, it doesn’t sharpen ideas. It sands them down. It rewards familiarity, penalizes risk, and quietly pushes everything toward the center. Safe. Polite. Forgettable.

That’s how brands spend millions and still disappear.

The brands people believe in don’t ask for permission. They don’t optimize their way into relevance. They decide what they stand for, build a world around it, and let the right people find them.

Liquid Death didn’t win by playing it safe and Apple didn’t win by consensus. They won by conviction.

Here’s the uncomfortable truth: If everyone likes your brand, no one loves it.

Love requires edges. It requires taste. It requires the courage to be misunderstood. Cult brands aren’t built by chasing approval, they’re built by expressing belief so clearly that the right people feel seen and everyone else self-selects out.

That’s not arrogance, it's leadership.

Metrics have a role, but not where most brands put them. Data belongs in distribution, not creation. Use it to amplify what you’ve made, not to decide what’s worth making.

Because algorithms don’t start movements. People do.

If you want attention, optimize. If you want devotion, decide.

Stop asking what the audience wants. Create from belief.

Stop chasing relevancy and start creating gravity (and gravity doesn’t ask for permission).

How to Respond Like a Career Politician

 

There’s a special dialect spoken in the marble hallways of power and the carpeted conference rooms of the C-suite. It’s fluent, confident, and utterly empty. 

It sounds smart. It feels responsible. It gives the comforting illusion that someone, somewhere, is in charge … while carefully saying absolutely nothing. 

This is the language of professional non-answers: a beautiful, aerodynamic form of bullshit designed to glide past accountability without ever landing on a real position. If you’ve ever listened to a politician or business executive talk for five minutes and realized you learned nothing, you’ve already heard it.

Below is a field guide to that language ... a greatest-hits list of phrases that masquerade as thoughtful leadership while doing the far more important job of protecting the speaker from risk, commitment, or reality. Memorize these and you too can sound wise, measured, and deeply engaged … without having to actually think, decide, or say anything at all.

"I'm not saying I'm for or against this - quite the opposite". 

 

"We need to have a serious conversation about this moving forward."

 

"I think we can all agree that this is a complex issue with valid concerns on both sides."

 

"At the end of the day, we have to circle back to our core values and leverage our strengths."

 

"Let me be clear: this isn't about choosing between A and B, it's about finding the right balance."

 

"We're committed to a holistic, 360-degree approach that takes all stakeholders into consideration."

 

"I'd caution against false choices here. The real question is how we move the needle on outcomes."

 

"Look, I think the data speaks for itself, and we need to let the process play out."

 

"This requires a nuanced approach that doesn't lend itself to soundbites or simple solutions."

 

"We're laser-focused on delivering results while ensuring we do this the right way."

 

"I'm not prepared to speculate, but what I can say is that we're exploring all available options."

 

"The bottom line is we need to be strategic and thoughtful as we navigate these headwinds."

 

"I hear what you're saying, and I want to be transparent: this is something we're actively monitoring."

 

"We're taking a data-driven approach while also listening to the voices that matter most."

 

"I think it's important we don't get ahead of ourselves here. We need to let the facts guide us."

 

"This is about building sustainable frameworks that create long-term value for everyone involved."

 

"We're cautiously optimistic, but we recognize there's still work to be done."

 

"I'm not going to litigate the past. What matters is where we go from here."

 

"We need to have guardrails in place while also not stifling innovation and agility."

 

"At this point in time, we're focused on aligning our priorities with stakeholder expectations."

 

"Let's not lose sight of the bigger picture while we're dealing with the tactical realities on the ground."

 

"I think reasonable people can disagree, but we all want the same thing at the end of the day."

 

"We're committed to transparency and accountability as we work through this process."

 

"This isn't a binary question … we need to thread the needle between competing imperatives."

 

"I want to be very careful not to prejudge the outcome, but we're cautiously encouraged by early indicators."