Thinking about the Bitter Lesson and the onset of advancing technological capabilities. It’s safe to say that AI has meaningfully impacted the Millennium Problems in mathematics. So it seems pragmatic to ask, “Is insert problem in insert domain harder than that?” It probably isn’t; act accordingly and manage the resulting risk exposure.

One way to do that is to make Pascal’s wager, but on moats instead of God: adopt the stance that technical moats don’t exist. Over a decade they erode, and something else must form a persistent differential advantage. Except we’re not dealing in decades now; we’re on Lenin time.