Not the End, But an End, Maybe

There is a not-so-subtle emotional rotation happening beneath the surface of markets. For the past two years, artificial intelligence has largely been a story of acceleration — capital spending cycles, semiconductor constraints, model improvements, hyperscale infrastructure. The dominant narrative has been about expansion and possibility. AI was framed almost exclusively as upside.

But markets rarely hold a single emotion for long.

Alongside the enthusiasm, another sentiment has emerged. Not doubt that AI will work, but concern about who it works for — and who it works against. As the practical applications become clearer, investors are beginning to ask harder second-order questions. What happens to traditional software layers? What happens to information intermediaries? What happens to portions of white-collar labor that were once considered insulated?

Every technological shift carries this duality. When photography arrived, it did not end art, but it did end portrait painting as it had existed. The automobile did not end transportation, but it did end the horse economy. The internet did not end commerce, but it permanently altered retail and distribution. Transformation rarely eliminates everything; it reshapes the structure.

This is not the end — but an end, maybe (credit for this line goes to Allen Levi, the author of Theo of Golden, a wonderful story).

That distinction matters for investors. Markets are forward-looking mechanisms. When a structural shift becomes visible, pricing adjusts before the income statements fully reflect it. Today, certain AI infrastructure companies are valued for extraordinary growth trajectories, while other segments are discounted under the assumption of margin compression or displacement. Exuberance and fear coexist, and both are priced simultaneously.

History suggests that in periods like this, markets tend to overshoot in both directions. Early beneficiaries are often priced as permanent monopolies. Potentially disrupted industries are sometimes priced as if adaptation is impossible. Reality typically lands somewhere in between. Productivity gains diffuse over time. The benefits broaden. New categories emerge that were not initially obvious.

A business owner recently framed the employment question in a way that feels grounded: “I can’t give an educated opinion on a major reduction in the labor force, but I can tell you that the employees who embrace AI tools will take the jobs of those who resist them.” That observation is practical rather than dramatic. It does not predict collapse. It predicts redistribution.

The same logic likely applies to companies. Businesses that integrate AI thoughtfully — using it to improve speed, insight, cost structure, and client experience — will gradually take share from those that ignore it. The technology itself is not the determinant of survival. Adoption is.

From an investment perspective, that nuance is important. We are not in the business of predicting which individual job titles disappear or which specific firms dominate the next decade. We are observing a broad productivity shift that will reward adaptability. Over time, AI’s influence will extend beyond semiconductors and data centers to logistics optimization, healthcare diagnostics, portfolio management, manufacturing efficiency, energy management, and small-business operations. Its economic impact is likely to diffuse, not remain concentrated.

That diffusion is part of why we think emotional extremes — both euphoric and apocalyptic — can create opportunity.

For us, this is not an abstract debate. We are not simply allocating to AI narratives; we are building with the tools ourselves as the cost and friction of developing applications drop precipitously. Over the past year, we have developed internal applications powered by artificial intelligence that allow us to operate more efficiently and with greater precision. These are not generic, off-the-shelf products. They are purpose-built systems that align with how we think about risk, research, and client service.

We use AI to accelerate research synthesis across macroeconomic inputs, enhance portfolio diagnostics and scenario mapping, streamline compliance workflows, and improve clarity and customization in our communication with advisers and clients. The technology does not replace judgment. It expands bandwidth. It reduces friction. It allows us to spend more time interpreting information and less time assembling it.

In that sense, AI has been less about cost-cutting and more about cognitive leverage. It improves how we execute, not just how we calculate.

This is why the fear narrative deserves perspective. AI will almost certainly change labor dynamics and margin structures in parts of the economy. Some business models built on information friction may compress. But adaptation has historically been more powerful than elimination. Railroads did not destroy economic activity; they rechanneled it. Electrification did not shrink industry; it expanded it. The internet did not reduce commerce; it multiplied it.

AI appears to be following that same arc.

For investors, the task is not to choose between enthusiasm and fear. It is to recognize where pricing has become unbalanced and where adaptability is underappreciated. Technological revolutions tend to reward participation and discipline while punishing narrow conviction. The early winners are not always the only winners, and the early losers are not always permanently impaired.

This is where diversification and process matter. Productivity shifts rarely remain confined to a single sector or theme. They spill outward. Margins change in one industry and improve in another. Costs compress here, and efficiencies expand there. The full economic effect is broader than the initial narrative.

Our framework is designed with that in mind. Rather than concentrating exposure around a single technological storyline, we focus on understanding the broader regime — the interaction between growth, inflation, credit conditions, and market leadership. When structural change accelerates, it often creates dispersion. Dispersion creates opportunity, but it also requires discipline.

We do not need to predict the exact winners and losers of AI to participate in its long-term impact. We need to remain diversified enough to capture its diffusion and flexible enough to adapt as leadership evolves. The goal is not to chase what is loudest, nor to hide from what is disruptive, but to stay positioned for the broader arc of productivity and economic adjustment.

This may not be the end of industries. But it is likely the end of certain ways of operating within them.

And history suggests that those who embrace the tools — whether individuals, businesses, or investors — tend to write the next chapter.