In 1900, New York City faced what many believed to be an existential urban crisis: horse manure. Hundreds of thousands of horses powered transportation, commerce, and daily life. The waste problem was so severe that newspapers and city planners openly debated whether cities could continue to function at scale. Entire industries emerged to manage, transport, and mitigate the issue. Capital flowed toward “solutions.”

Within a few years, the problem vanished.

Not because it was solved — but because it was displaced. The automobile didn’t improve manure management. It made horses irrelevant.

This pattern repeats throughout history. The most important problems are rarely solved head-on. They are leapfrogged by paradigm shifts that redefine the system entirely.

That distinction matters deeply for investors.

Displacement vs. Solution

Many businesses are built around addressing a clearly defined pain point: inefficiency, cost, complexity, friction. In stable systems, this can work for decades. But during periods of technological acceleration, the risk is not that a company fails to execute — it’s that the problem itself stops mattering.

We’ve seen this dynamic repeatedly in recent years.

Developer tooling is a useful example. Platforms designed to democratize coding, simplify development environments, or abstract infrastructure felt compelling when software creation was bottlenecked by setup complexity and technical skill. But as large-scale AI coding assistants became embedded directly into operating systems and productivity suites, the problem definition shifted. The constraint was no longer who can code — it became what should be built at all.

What once looked like a durable wedge suddenly became redundant.

This is not a critique of execution. It’s a reminder that technology doesn’t just compete — it reframes.

A Modern Analogy (Not a Perfect Overlay, but the Same Physics)

No historical analogy maps perfectly onto modern technology. Markets are more complex, capital moves faster, and platforms are more interconnected. But the underlying physics of displacement remain the same.

A useful contemporary example is developer tooling.

Platforms like Replit emerged with a compelling promise: lower the barrier to entry for software creation by abstracting environment setup, infrastructure, and complexity. At the time, this addressed a real bottleneck. Coding required specialized knowledge, local configuration, and nontrivial setup. The problem was clear, and the solution was elegant.

What changed was not execution — it was the layer at which the problem was solved.

As AI-powered coding assistants became natively embedded into dominant platforms and workflows — particularly within large productivity ecosystems like Microsoft’s — the constraint shifted. The question stopped being “How do I make coding easier?” and became “Do I even need to write this code myself?”

This is not a clean one-to-one displacement. Replit and similar platforms still offer value in specific contexts. But the center of gravity moved. What was once a standalone solution became a feature. What was once a platform-level wedge risked being absorbed upstream.

That is the critical lesson.

The danger is rarely that a company builds the wrong product. The danger is that a much larger system reframes the problem in a way that bypasses the original solution entirely.

This is the same dynamic that has repeated throughout technological history: when platforms absorb functionality, entire categories compress — not because they failed, but because they were solving yesterday’s version of the problem.

Exponential Change Breaks Linear Thinking

One of the most underappreciated aspects of modern investing is how poorly human intuition handles exponential systems. Progress feels slow — until it isn’t. By the time change becomes obvious, it has already compounded beyond control.

AI sits squarely in this category.

At any given moment, the technology can feel incomplete: laggy, expensive, prone to hallucination, or narrowly scoped. That leads many observers to discount its long-term impact. But exponential systems don’t need to be perfect to be inevitable. If a tool is 70% useful today, it doesn’t need heroics to reach 90% — it just needs time.

And unlike prior technology waves, AI improvement is recursive. The system improves itself.

This has profound implications for capital allocation.

The VC Monoculture Problem

One of the clearest signals of displacement risk today is the growing homogeneity of venture deal flow. Increasingly, every company is presented with an AI narrative. Businesses that had no explicit AI angle two years ago now lead with it in pitch decks and annual meetings. Language models, automation, agents — the framing is everywhere.

The uncomfortable question is not whether AI is important. It is.

The real question is: where did everything else go?

If nearly all capital is chasing AI-branded opportunities, then either:

1. Entire categories of businesses have quietly disappeared (unlikely), or

2. They are being systematically underfunded (potential opportunity), or

3. Capital is clustering around a narrative faster than fundamentals can justify (bubble risk)

History suggests that when capital crowds into a single theme, dispersion collapses. In those environments, even strong companies can suffer when sentiment reverses. The tide doesn’t discriminate.

Moats Under Pressure

Traditional investment frameworks emphasize durable moats: scale, switching costs, capital intensity, regulatory barriers. These remain important — but they are not immune to displacement.

AI introduces a new kind of threat: not competition from a better product, but elimination of the need for the product at all.

Consider energy and compute. Today, massive investment is flowing toward data center infrastructure, cooling, power generation, and grid expansion. These may be necessary — but they also assume current architectures persist. If algorithmic efficiency improves faster than hardware demand, the constraint shifts again.

Sometimes the better mousetrap never arrives — the mouse disappears.

How We Think About It

This doesn’t mean avoiding technology. It means respecting second-order effects.

When we evaluate opportunities, we spend less time asking:“Is this a good solution?”

And more time asking:“What happens if the problem definition changes?”

We are drawn to businesses with:

• Structural resilience

• Low dependence on a single technological assumption

• Economic models that benefit from change rather than fight it

• Cash flow and optionality, not just narrative momentum

In periods of rapid displacement, humility matters. The goal is not to predict the future perfectly — it is to avoid anchoring capital to problems that may not exist tomorrow.

Closing: Investing for the Right Layer

The horse manure crisis was real. The capital deployed to solve it wasn’t irrational.

It was just aimed at the wrong layer of the system.

Periods of technological acceleration reward investors who think less about answers and more about assumptions. The most durable opportunities tend to sit above specific implementations — aligned with incentives, economics, and adaptability — rather than tightly coupled to a single version of the world.

When technology moves slowly, optimization wins. When technology moves exponentially, positioning wins.

Our job is not to solve today’s problems faster than everyone else.

It is to recognize which problems may quietly disappear — and to allocate capital accordingly.