The renewed rise of the defense industry is another sign that today is starting to resemble the 1930s. Germany’s push to convert railcar plants, auto suppliers, and even food-tech firms into military producers shows how quickly industry is adjusting. Markets have already priced this shift in. But this is also an old model of defense economics: heavy equipment, long contracts, and government-backed demand designed to keep factories running, not to create fast-growing companies.

Europe’s broader increase in military spending follows the same logic. It prioritizes industrial capacity and security of supply over innovation or scale-driven returns. The U.S. is moving in a different direction. Capital is flowing to defense-tech companies like Anduril and Palantir, which are valued more like software firms than traditional defense contractors. Investors are betting on speed, data, and autonomy rather than steel and assembly lines.

The war in Ukraine makes the divide clear. Modern warfare is becoming faster, more automated, and less dependent on large human forces. Germany’s industrial shift will protect jobs and ensure steady revenue, but the real upside is going to companies that own software, sensors, and rapid deployment systems. Whether Europe ends up with global defense leaders or just fully booked factories will depend on whether it can close that gap.

Economists point to high interest rates, trade tariffs, and economic uncertainty as reasons unemployment is rising. These pressures are pushing companies to cut costs through hiring freezes and layoffs. That part is familiar. What is different this time is that generative AI is directly shaping employment decisions.

AI is not a side factor. It is a core driver of how work is being reorganized. Companies that once might have waited for conditions to improve are now redesigning workflows, flattening management, and delaying rehiring because fewer people are needed to do the same work. This is why the current slowdown looks different from past cycles. Many roles being cut or paused, especially in professional services, finance, media, and tech, are not coming back. They are being eliminated or merged into fewer jobs with higher output expectations.

As a result, unemployment pressure is likely to grow even if growth stays positive and inflation continues to ease. Job searches will take longer, and white-collar layoffs will spread beyond tech. In the coming quarters, unemployment not inflation is likely to become the main economic and political issue. Prices are cooling, but job insecurity is rising, particularly among educated workers. Policymakers risk fighting the last battle as the labor market adjusts to a new link between growth, productivity, and employment.

I think about this subject a lot because I have two sons, 12 and 10, and I’m watching choices take shape in real time. I’ve always been a big believer in being a generalist, and that belief has only strengthened in the AI era. Early specialization often looks impressive. Kids rack up wins, rankings, or credentials fast. But those gains usually reflect timing, narrow training, or structural advantage, not durable ability. Research across sports, education, and cognitive development shows the same pattern. Systems optimized for early performance tend to trade long-term upside for short-term validation. Burnout, plateauing, and disengagement show up later.

The better path is generalist first, specialist later. This is the logic behind the centaur model. Early years should focus on building range: exposure to multiple disciplines, sports, and ways of thinking, alongside human skills that machines cannot replace empathy, judgment, intuition, and ingenuity. Depth should come later, chosen deliberately. In an AI-augmented world, success won’t come from knowing one narrow thing better than everyone else. It will come from connecting ideas across fields and using AI as leverage, not a crutch. Generalists compound over time because they see patterns earlier, adapt faster, and sustain motivation longer.