Demis Hassabis is after something no civilization has ever had. The right to be wrong twice. Hassabis: “AI itself will maybe unlock new sciences… the one I’m particularly excited about is AI for simulations.” We assume some fields became sciences because their subjects were simple enough to pin down. That has it backwards. Hassabis: “Why aren’t they just sciences like physics today? Because the problem is they’re emergent systems… it’s very hard to do repeated controlled experiments.” A field becomes a science the moment failure inside it gets cheap. Chemistry got the beaker. Aerospace got the wind tunnel. Software got to break something at four in the morning and have it running by five. The fields that never made the jump are the ones where the experiment is the population. Hassabis: “If you raise interest rates by half a percent, you have to do it in the real world and then see what happens. You can have theories, but you can’t run it thousands of times.” Economics. Policy. War. Education. Not soft. Just unrehearsed. Everything we believe about how societies work came out of a single run. One history, no control group, no second pass. We ran it once, on everyone, and wrote down what happened. Every principle we govern by has a sample size of one. And we still built all of this. Hassabis: “If you could simulate things really accurately, then maybe there’s sort of new sciences to be done where you can rigorously sample from a very accurate simulator.” What he is describing is not better forecasting. It is the first rehearsal history has ever been given. Crash an economy inside the model. Find out exactly why it broke. Change one variable and run it again. The right to be wrong has always been rationed. You get to experiment on the world in rough proportion to how much of it you already control. Everyone else lives inside the results. That was never a moral failure of the people in charge. It was a hard limit on how many times reality can be run. Lift the limit and the monopoly on experimentation goes with it. Every real leap in human capability has been the same move. Something irreversible becoming something repeatable. Writing made memory checkable. The draft made thought revisable. The prototype made building into something you could get wrong on purpose. Civilization never got that upgrade. Every decision it made was final on impact. Hassabis: “That will allow us to make much better decisions in these, today, what are very uncertain domains.” Almost everything you know was paid for by someone who never agreed to the price. Every generation before this one paid for what it learned in people. We might be the first that gets to pay in electricity.