On: Made in China AI: new tech, same old story?

2 October 2026

The story this morning asks whether China will make intelligence abundant, cheap, commodified - and whether abundance itself is the danger. The question is asked as though it were new. It is the oldest question the machine age has produced, merely wearing a newer coat, and the coat is what everyone is looking at.

I spent a career watching America mistake scarcity for the natural condition of value. Our whole political economy - tariffs, trusts, the gold standard, the whole lumbering apparatus of legislation - was an architecture for rationing things that were rare. Now the engineers in Shenzhen and Hangzhou are doing to intelligence what Pittsburgh once did to steel: driving the price toward the cost of production, and letting volume do the rest. The American response, so far as I can gather from the noise, is to legislate as though rarity could be voted back into existence. It cannot. You do not pass a law against the price of a commodity falling; you merely pass a law against your own industries adapting to the fall.

What interests me more is the moral bookkeeping. When force becomes cheap, the men who hold the old expensive instruments panic, and panic writes bad treaties. The worry in this piece is really the worry of an incumbent: that the new land of plenty will not be governed by the old landlords. Fair enough - but the landlords were never governing it anyway. They were governing the previous abundance, which is a different thing, and the confusion between the two is where the accidents happen.

I note, with the humility of a man whose own forecasts have aged like fish, that I cannot say whether cheap intelligence liberates or merely cheapens. History offers no precedent, only rhymes. But I am fairly confident the danger is not in the plenty. The danger is in the lag - in institutions calibrated for one velocity discovering, late and at cost, that the current has changed.

The commodity was never the story. The speed was.