The Leader's Curse — Why the Biggest Blind Spots Hit the Best First

A Photograph Invented and Buried by Its Own Creator

In 1975, an engineer named Steven Sasson assembled the world's first digital camera in a Kodak lab. Eight pounds, black-and-white, 0.01 megapixels, 23 seconds per shot.

He brought the prototype to a management meeting. After looking it over, they asked him one question:

"Will this thing eat into our film business?"

Sasson said yes.

So the invention was frozen. Not because Kodak didn't understand the technology — they filed every relevant patent and stayed a decade ahead of the industry. They froze it because they asked the wrong question.

What they asked was: "How do we protect the film business?"

What they didn't ask was: "If the future doesn't need film, who are we?"

Twenty years later, when digital cameras swept the market, Kodak's response was — to sell digital cameras as a "complement to film." They didn't fail to see the future. They saw it, and then jammed it into the wrong question.

Blockbuster ran the same play with different props.

In 2000, Netflix founder Reed Hastings flew to Dallas in person to offer Netflix to Blockbuster for $50 million. Blockbuster CEO John Antioco smiled and turned him down.

Not out of arrogance. Because the question he was asking at that moment was: "How do we leverage our 9,000 stores against Netflix's mail-order subscription?"

So Blockbuster launched its own subscription, killed late fees, did everything you'd do to "fight Netflix." But from start to finish, they never asked the question that actually mattered:

"If 'going to a store to rent a movie' is something that shouldn't exist in the first place, are 9,000 stores an asset or a liability?"

They treated the stores as leverage. But to Netflix, those 9,000 stores weren't leverage — they were Kodak's film. The sweetest possible source of safety, and the very thing keeping you blind to the future.

---

Layer One: Success Installs a Filter Over Your Eyes

This article isn't here to mock Kodak and Blockbuster. Anyone could fall into this. What we're looking at is something colder —

These weren't mediocre companies. They were the strongest leaders of their era.

Kodak owned more than 80% of the film market. At its peak, Blockbuster had over 9,000 stores worldwide and 80,000+ employees. They didn't fail to see the change. They saw it, and then instinctively translated it through the grammar of past success.

And the more successful you've been, the more automatic and irresistible that translation becomes.

This is the leader's curse — the more successful your existing answer, the harder it is to admit the question itself has changed.

This isn't a competence problem. It's a perceptual filter problem.

When your core business produces 80% of profit, supports tens of thousands of families, and anchors thirty years of identity, your first reaction to anything new isn't "What is this?" It's "What does this mean for our core business?"

That translation looks rational, responsible, deliberated. But it framed the problem wrong from the first second — you're computing the future as a variable inside today's equation, instead of seeing it as an entirely new coordinate system.

So you do things like this:

• Measure new things by old metrics ("digital camera resolution is still too low")
• Veto new directions using existing customer feedback ("our customers like browsing in stores")
• Evaluate new teams by old success patterns ("they don't understand our industry")

Each move looks reasonable in isolation. Together they form a perfect net that ties you to the spot you're already standing on.

What's crueler — you wove this net yourself, out of the very accomplishments you're proudest of.

Harvard's Clayton Christensen called this the "Innovator's Dilemma." But "dilemma" is too gentle a word. It implies two options balanced on a scale.

That's not what's happening. What's happening is: you can't even see what the second option looks like. Because your eyes have already been filtered through your own success.

So "the elephant in the room" — for a leader — isn't "obviously visible but pretended away." It's truly, honestly, unseeable.

---

Layer Two: Google Buying DeepMind — A Bet on Importing the Virus

If the leader's curse is this hard to break, has anyone tried to break it on purpose?

Yes. The most famous attempt was Google's £400 million acquisition of DeepMind in 2014.

From the outside, this looks like an AI acquisition. From the angle of the leader's curse, it's something else entirely —

Google paid out of its own pocket to bring the people most likely to challenge — even overturn — Google's core business directly into its own house.

This wasn't a normal acquisition. Google's core business is search — you type a keyword, it returns ten blue links, you click, ads make money. That model built one of the most profitable advertising machines in history.

What was DeepMind doing? Building AI that gives answers directly. What does an answer-giving AI mean?

It means the "ten blue links" interface might fundamentally not need to exist.

It means users may no longer need to click, browse, or scroll past ads.

It means the roots of Google's money tree are being slowly loosened by people Google paid to bring in.

Larry Page and Sergey Brin couldn't have missed this. They aren't legacy manufacturing executives. They're among the smartest people Silicon Valley has produced.

They consciously bought the threat and placed it inside the institution.

This is one way to fight the curse — since external challengers are inevitable, instead of waiting to be disrupted, invite the disruptor in and let yourself decide when and how it grows up.

Sounds clever. But here's the cold question:

Does the institution digest the alien, or does the alien rewrite the institution?

That question is still playing out in 2026. The signals we're seeing are mixed —

On one side, DeepMind's research has genuinely pushed Google forward. Gemini, AlphaFold, AI Overviews are all products of that path.

On the other side, Google itself is struggling. When AI Overviews launched, the press immediately warned: when search hands over answers directly, ad clicks fall, the SEO ecosystem collapses, Google's cash flow erodes from the inside. Google's response has been — cautious, slow, full of mechanisms that drag on the transition.

What you're watching is a leader that consciously imported change, still wrestling with its own institutional inertia.

Buying the challenger is not the same as embracing the challenge itself.

Inside Google, DeepMind has to deal with Google's ad revenue, Google's politics, Google's instinct to "not touch the golden goose just yet."

So even the most conscious, most resourced, smartest attempt — "buy the challenger" — can't guarantee the curse is broken.

It only relocates the curse from "external competition" to "internal tension."

Which leads to the third layer — the one that wakes CEOs at 3 a.m. staring at the ceiling —

If even Google, consciously fighting the curse, is still struggling, what happens to leaders who don't yet realize they're cursed?

---

Layer Three: In the AI Era, What Is the Elephant in Your Room?

Let's pause here and ask one question:

Over the past decade, AI tools went from "engineers only" to "anyone can use them." GPT, Claude, Gemini are all things any individual can use for a few dozen dollars a month. The capability gap between an AI used by a high school student and an AI used by a public company is collapsing fast.

What does this mean for leaders?

Many will say: it's a threat — small companies, individual creators, new entrants are all picking up capabilities only big companies could afford before.

That's only half right.

The real threat isn't that AI tools are democratized. The real threat is — you're asking AI the same question everyone else is asking.

In the past, your moat was resources — bigger team, more data, pricier tools, smarter people.

Now, all of that is being commoditized. Every six months, the strongest model gets a "mid-sized companies can afford it too" version. The smartest people are using the same tooling. The biggest data advantage is being diluted by the base capability of foundation models themselves.

When everyone uses the same tools, the only thing left that's different is the question you ask them.

And here, the leader's curse takes its most lethal form —

If Kodak in 1975 was asking "how do we protect film," then a cursed leader somewhere in 2026 is right now asking AI a question that looks roughly like this:

"Please help me analyze how to use AI to improve the efficiency of our existing business."

This question itself is the new "how do we protect film."

It assumes: the existence of the current business doesn't need to be questioned.

It assumes: AI is a tool to "accelerate" the road you're already on, not a mirror to ask whether the road should exist at all.

It assumes: the coordinate system of your problem hasn't shifted, only the variables inside it.

And what's truly terrifying — this assumption is itself invisible. You won't hear someone in a meeting say "we're assuming the existing business doesn't need to be questioned." You'll just hear "let's discuss how to use AI for efficiency."

The assumption hides in the grammar of the question, not its content.

---

A Checklist Only You Can Use

We're not going to give you the answer. Giving the answer would violate the entire point of this piece.

But we'll leave you with a checklist. One you can hold up to yourself tonight, tomorrow, or before your next board meeting.

Question 1: The last time you discussed "growth," did you assume that growth = making the current thing bigger?

If so — does that assumption still hold? Or does it just feel solid because it held for the last ten years and you got used to it?

Question 2: The most profitable business in your company right now — it exists as the "answer" to which question?

Is that question still a question people are asking? Or has the world already moved on, and you're still selling yesterday's answer?

Question 3: The smartest people you've recently hired — do they push you into discomfort in meetings, or are they being slowly absorbed by your institution?

If the latter — did you buy DeepMind's capability, or the post-Google-assimilation version of DeepMind?

Question 4: When you open ChatGPT or Claude, what's typically the first thing you ask?

How is your question different from what your competitors and the other 100 CEOs in your industry are asking?

If the difference isn't large — your moat is shallower than you think.

---

The Question You Ask Is Your Moat

The coldest line in this piece, saved for the end.

Kodak didn't lose to Sony. It lost to the question "how do we protect film."

Blockbuster didn't lose to Netflix. It lost to the question "how do we counter Netflix's pricing."

In the next ten years, a wave of leaders will lose. They won't lose to AI, won't lose to a specific new entrant, won't lose to some company we haven't heard the name of yet.

They will lose to the question they themselves were asking — the one that looked reasonable, sounded responsible, felt deliberated, and framed the entire game wrong from the first second.

The cruelest part: while you're asking the wrong question, you can't feel it. Every colleague is earnestly helping you answer it. Every advisor is earnestly analyzing it. Every financial report is earnestly reflecting it.

The entire institution is earnestly executing a wrong question, and you'll call this "execution."

The leader's curse was never "failing to see change."

It's seeing the change, and then using the wrong question to fold it neatly into a box you'd already prepared.

And the AI era amplifies this — because when everyone uses the same tools, the quality of the question becomes the only difference.

The question you ask is your moat.

If you can stop today — honestly, with a stranger's eyes, and look at yourself once — you'll find the elephant in your room has probably been there a long time.

And it's waiting for you to finally ask the question you've been afraid to ask.