When I am in Max, I am spending insane amounts of tokens for relatively small tasks that are surely not taking up that much spend or context.
Surely, there could be a way to optimize how much Max spends, maybe by dynamically making the context window adjust based on calculated need we could spend less.