Simulation efficiency systems often look general but only wire one lever

When a sim has an aspect->multiplier efficiency system, dont assume every aspect actually bites — grep for the consumption call site of each aspect. In practice an org_efficiency(aspect) system had only the storage aspect read anywhere; a production or transport aspect was inert framework until code in the tick loop actually multiplied by it. Also, when making research grant an operable output, prefer the conservation-safe lever: reducing a process INPUT per cycle (finite stock lasts longer -> more output over time) rather than a facility that produces a resource from nothing, which is an unbounded faucet. Gate the whole thing behind an opt-in meta flag so existing runs stay byte-identical.