In reply to Venkatesh Rao's Blyg — Venkatesh Rao
Sorry for more Chapman-channeling, but this is exactly what his situated action work was set out to disprove: planning doesn't work even in fully legible domains, because of computational complexity issues. I assume dead-reckoning is what AI people call planning (figuring out a set of actions to achieve a goal, ahead of actually performing any).
Let's say: "fully legible" ≡ "perfectly playable domains" ≡ "simulation"
Whether his arguments apply to planning as it exists in the real world is another matter.
For later: people are gravitating to a dead-reckoning model of intelligence by treating length of autonomous runs as a kpi. I think this is ill-posed. Perfectly playable domains allow dead-reckoning to substitute for situated orienting and maneuvering because simulation is 100% fidelity and tractable. Where that’s not true it’s the wrong mental model.