Taste is a muscle, not a gift — train it by forecasting every result before you see it

Vivek reframes taste from innate gift to trainable faculty: “predict the result of every experiment before you run it. cover a paper’s results section and guess the numbers from the method alone. mark down which of this month’s releases will matter in two years and check your hit rate later.” The training signal is explicit: “a forecast plus a correction, repeated a few hundred times, is how every good model gets trained, including the one in your head.” Taste is just a well-calibrated predictor, and calibration comes from logged prediction-error, not exposure.

This is the human-judgment version of why AI compresses the distance between idea and execution but not between good and bad judgment — execution gets cheap, but the forecasting muscle that judges what to build is exactly what doesn’t, and deliberately training it is how you avoid being on the wrong side of Amplification widens the judgment gap — AI magnifies clear thinking into compounding advantage and confused thinking into accelerating waste. As tools dissolve constraints and When production constraints dissolve, the bottleneck shifts from execution to judgment, the forecast-then-correct loop is the cheapest way to grow the bottleneck faculty. It’s also the upstream skill for Reason backward from an outcome you want to exist — it manufactures originality that absorbed problems can't: you can only reason backward toward an outcome worth wanting if your predictions about what will matter are calibrated.