My mother taught literature and lived what she taught, and she made me look at the life, not only the lesson. She handed me a computer she could not operate.
I have a hard-science degree and a math minor. I'm going to make the liberal arts argument anyway, because I learned it by watching.
The point of a good education is not the fields. It is the connections between them. Nothing, wherever it came from, is alone.
I learned that young. In ninth grade I skipped lunch to sit in the school library, and in one year I read Les Misérables, the Inferno, Plato's Apology and Aristotle. None of them was easy alone. Together they explained each other.
Here is what that looks like at work. An AI rack weighs about 3,000 pounds and draws about 120 kilowatts. Hold those two facts, structure and electricity, and a third appears: the floor. Then the cooling, the shipyard built to carry ships in a cold place, the people who can pipe it, the politics of a state pausing large data centers. I grew up next to the ocean. I sail the Delaware and the coast from Atlantic City to Block Island and Martha's Vineyard, and I have walked the shipyards at Bath, Philadelphia and Norfolk. I know water, and I know those yards, by work and not abstraction. No single field asks whether a shipyard town could hold an AI data center. Someone fluent on both sides of each seam can.
That is why an engineer should argue for breadth. Abstraction hides the layer beneath it, and it fails at the boundary between fields. For an investor, the weak assumption in a plan usually sits at a seam where no one on the team is fluent on both sides.
Knowing is also a position. My great-grandmother's brother was good with numbers. He worked out that the man who owned the land had been shorting the families who worked it, and he presented the figures. In that place and decade, that was a death sentence for a black man. He was run out. He went north and ran an oil-delivery business. Poor white families were caught in the same ledgers, for different reasons. I know why the caged bird sings, from more than one angle.
Breadth runs fast and loose. Thirty years after reading a book, I keep the feeling. The particulars are work. While drafting this I credited a famous book to the wrong author: right idea, wrong name. The discipline is checking the label before you publish it.
Nothing is alone. Show up. Look around. Learn what is necessary.
I run Osparna, a diligence firm that also makes early investments, and I run large AI models on my own hardware. This was drafted with Claude, made by Anthropic; where I say "I," the judgment is mine, not a source's.