"The Nielsen numbers are telling us the wrong story. When a show gets 800,000 viewers who'd genuinely rearrange their week around it, that's not a failure. That's a constituency. The industry doesn't have an audience problem. It has a metric problem — and the metric problem is upstream of every creative decision that follows."

"What the old studio heads did with intuition and lunch meetings, the recommendation engine does with watch-time data and completion rates. Both are wrong about the same percentage of the time. The difference is that one of them could be argued with, fired, or embarrassed at a festival. The other one just updates silently at 3 a.m."

"We spent fifteen years calling television cinematic as if that were the highest compliment we could pay it. What we meant was: it looked expensive. What we missed was that the best television was doing something film never could — living inside a character's Tuesday for sixty hours. The bar didn't get higher. It got different. We're still arguing about height."