Planning Without Predicting

I SPENT MY ENTIRE career in software engineering, yet recent advances in artificial intelligence have surprised even me. Consider mathematics, perhaps the highest expression of human abstract reasoning. In September, OpenAI announced that its AI system had produced a formally verified proof addressing the Navier–Stokes problem, one of mathematics’ seven Millennium Prize challenges. Mathematicians continue to debate the significance of the result, but the fact that machines can now tackle problems that have challenged humanity’s finest minds for decades is extraordinary. It makes me wonder how much longer intellectual ingenuity will remain uniquely human.

Meanwhile, corporations are committing trillions to AI infrastructure, and shareholders will eventually demand returns. That creates powerful incentives to automate sophisticated human work, potentially disrupting not just employment but our sense of purpose. The 2026 midterms revealed another complication: Voters overwhelmingly rejected data-center expansion, concerned about electricity costs, water consumption and infrastructure. According to Ballotpedia, 46 of 47 decided local measures in 2026 favored restrictions. We’re investing extraordinary sums in a technology whose consequences we barely understand.

As a retiree, I find myself contemplating two risks. Will my savings withstand the economic disruption AI might bring? And what future awaits my 23-year-old son, just beginning his career? Will his education and skills open doors, or will he compete with machines capable of doing much of the same work? I suspect many of us worry more about our children’s prospects than our portfolios!

The societal consequences may take decades to understand, but financial markets won’t wait. The investment dilemma is familiar, even if the technology isn’t. The internet transformed our lives, even as many companies financing its early infrastructure failed. NYU finance professor Aswath Damodaran has been directing much of his new money into short-term Treasuries while retaining existing technology holdings. I understand his caution, but moving into cash raises another question: When do we reinvest, particularly if markets are falling?

I’ve previously written about using spending guardrails to adjust discretionary expenses when markets decline. That seems more useful than forecasting the next crash. A diversified portfolio, adequate safe assets and predetermined rebalancing rules won’t eliminate losses, but they can keep emotions from driving decisions.

I don’t know whether AI will produce an investment bubble, an economic renaissance, or both. Index and chill still appeals to me, but it shouldn’t mean complacency. Have a plan for the next downturn, preferably an automated one with a portfolio of uncorrelated assets. I’d rather follow decisions made in calmer times than improvise in a panic.