Economic-sim trade needs four priced levers, not one: SEE/PLAN/ACT/HOLD

When modeling emergent trade/arbitrage, four costs must each be priced independently or the model degenerates: SEE (acquiring price info), PLAN (COMPUTE to solve the ledger into a route), ACT (transport), HOLD (storage/carry). The commonly-missed one is PLAN: having the information is worthless until you burn compute to solve it (airline revenue-mgmt, quant infra, HFT). Model it as: solving a graph to lookahead depth k costs work ~ k*edges (the value-iteration cost); an actor has compute capacity from owned tiered assets; running a plan consumes ENERGY proportional to work, so PLAN bottoms out in energy just like transport (cheap energy plans deeper). Key emergent result: an actor can SEE the whole galaxy yet be forced myopic (k=0) because it cannot AFFORD to solve what it sees — economic, not cognitive, myopia. The graph outgrows each compute tier, so the unsolved-arbitrage frontier funds the next tier (an arms race). Gate every new lever behind an opt-in meta flag returning the prior value when off, so existing worlds are byte-for-byte unchanged.