We’ve been taught to tell simple narratives. They are catchy and memorable. Let’s be honest. They work. Good narratives compress complexity into clarity and conviction. They give founders the language for their vision, recruiting, and fundraising. They give investors a shared vocabulary for pattern recognition. When Sam Altman or Dario Amodei framed their company’s mission to build AGI, or when Ben Thompson described aggregation theory, these were narrative compressions that unlocked real insight.

In the world of technology, where complexity is the norm, simple narratives are both seductive and dangerous. They promise clarity in environments filled with ambiguity. While they can guide action and unify thinking, they often obscure more than they reveal.

Daniel Kahneman called this “attribute substitution.” When we face a complex judgment, we unconsciously replace it with a simpler one. Simple narratives are attribute substitution at an industrial scale. We replace the hard question (“What is actually happening in this market?”) with the easy one (“Does this fit the narrative?”). We simply don’t let facts get in the way of a good story.

Consider a few narratives that shaped and misshaped technology investing:

• Winner takes all. In some markets, such as search and social networking, this proved largely correct, but enterprise software proved stubbornly multi-vendor. E-commerce never consolidated the way the narrative predicted. Even in cloud infrastructure, the oligopoly of AWS, Azure, and GCP defied the single-winner thesis. The narrative was a useful heuristic. Founders and investors who treated it as a law made expensive mistakes.
• First mover advantage. Google was not the first search engine. Facebook was not the first social network. The iPhone was not the first smartphone. The company that finds product-market fit in the right window wins. But “timing and execution matter more than sequence” is a harder story to tell than “be first.”
• AI will replace [x]. Today’s dominant narrative is directionally correct but operationally misleading. The simple version, that AI replaces humans in a neat, linear substitution, misses the more investable reality. Augmentation, new workflows, and entirely new categories of work tend to emerge alongside displacement. The companies building for the nuanced version of this future look very different from those building for the simple version.

The greatest danger of simple narratives is not that they’re wrong at the start. It’s that they become unfalsifiable over time. Thomas Kuhn described this in The Structure of Scientific Revolutions. Practitioners of a dominant paradigm don’t reject it when anomalies appear. They explain the anomalies away and carry on until the weight of contradictions becomes unbearable.

I lived this myself. In 1997, I declared that Amazon would kill Walmart. Today, Walmart is 30 times larger than it was 30 years ago. With each quarter of declining mall traffic and each confirmed brick-and-mortar bankruptcy, the thesis held true. This was confirmation bias at work. The world was messier than the story. E-commerce companies also failed. Customer acquisition costs online kept rising. Certain categories had persistent try-before-you-buy dynamics. Physical presence created brand equity that digital alone could not. Those who treated the simple narrative as a settled truth missed the omnichannel reality that ultimately prevailed.

None of this is an argument against narrative itself. We cannot operate without mental models. The argument is against attachment to narrative and against the moment when a useful simplification hardens into an unexamined belief.

Clayton Christensen understood this. His theory of disruption in The Innovator’s Dilemma was itself a narrative, and he spent the latter part of his career refining it, adding caveats, and drawing boundaries around where it applied and where it didn’t. He knew his framework was a map, not the territory, and he kept updating the map.

The best founders and investors hold their narratives loosely. They have a deep conviction about where the world is going but remain genuinely flexible about how it gets there. Tony Hsieh did this at Zappos. The narrative was about delivering happiness. The operational reality was a constant evolution of what that meant. Free shipping. Free returns. A 365-day return policy. A culture-first hiring philosophy. The narrative was stable. The strategy underneath it was always in motion.

Stuart Kauffman introduced the concept of “the adjacent possible,” the set of first-order changes available from any given state. Simple narratives tend to describe a distant possible while ignoring the adjacent one. “AI will replace all knowledge work” is a distant possible. “AI will first augment the workflows where data is structured, stakes are moderate, and feedback loops are fast” is an adjacent possible. The distant version is more exciting. The adjacent version is more buildable and more investable.

Our job is to operate in the adjacent possible while keeping the distant possible in peripheral vision. That requires resisting the gravitational pull of the clean story and doing the harder work of understanding what is actually true right now.

What narrative are you holding that feels so obviously true you’ve stopped testing it?

Every one of us has one. Find yours, and hold it as a hypothesis rather than a conclusion. Keep asking: What would have to be true for this story to be wrong?