Aaron Levie argues open vs closed barely moves AI economics, because the money was always in inference: "I generally think that the moneymaker in AI is inference. Ultimately the dollars are gonna flow to the infrastructure stack one way or another." "In a world where we're gonna have three to five players in the US that consider it existential to have leading models, you have enough of a competitive dynamic that the cost of tokens converges closer and closer to the cost of the infrastructure…