Signature arguments.

Every industrial revolution has been accompanied by a moral panic about employment. And every one has been followed by a larger, more complex economy with more work to be done — just different work. The AI revolution will be no different in outcome, but it will be unprecedented in speed.

The firms that thrive will not be those that automate the most. They will be those that redesign work itself — decomposing roles into tasks, understanding which tasks benefit from human judgment, and building organisations that can continuously reconfigure as the technology improves.

For every doubling of installed solar capacity, the cost of solar panels drops by roughly 20-25%. This learning curve has been remarkably stable for four decades — and yet forecasters, including the International Energy Agency, have consistently underestimated its implications.

The reason is simple: human intuition is linear. We expect next year to look roughly like this year, plus or minus a few percent. But exponential curves don't work that way. Solar energy is now the cheapest source of electricity in history in most of the world, and the cost continues to fall.

The semiconductor is the new oil. Control over chip fabrication, rare earth minerals, and AI training infrastructure now shapes geopolitical power as decisively as control over fossil fuels did in the twentieth century. Yet most corporate strategy still treats geopolitics as an external risk to be monitored, not a structural force to be incorporated into core planning.

The US-China technology decoupling is not a temporary trade dispute. It is a structural realignment of the global technology ecosystem into competing spheres of influence. Companies that depend on supply chains spanning both spheres face a fundamental strategic choice — one that will define their competitive position for decades.

In 2021, I published *Exponential* to articulate a pattern I'd observed across two decades of building, investing in, and analyzing technology companies: the pace of technological change has decoupled from the pace of institutional adaptation. I call this the exponential gap.

The gap manifests everywhere. Regulators struggle to govern technologies that evolve faster than legislative cycles. Corporate boards make three-year plans in industries being reshaped on eighteen-month timelines. Educational institutions prepare students for jobs that may not exist by graduation.

Financial services firms were among the earliest adopters of machine learning — in fraud detection, algorithmic trading, and credit scoring. But generative AI represents a qualitative shift. It doesn't just automate existing workflows; it makes entirely new ones possible.

The investment firms that will outperform are not those deploying AI to write research reports faster. They are those rethinking what a research analyst does when AI can synthesise every public filing, earnings call, and market signal in seconds. The role shifts from information gathering to judgment, pattern recognition across domains, and the kind of creative synthesis that AI cannot yet replicate.