Information You Can't Revise
Every intake form is a bet about when you'll know enough to stop asking questions. Most of the time the bet lands on the side of more, because more information reads as more rigor. What it actually buys is a longer runway before you commit.
There's a difference between information you screen with and information you steer with. Screening happens once: you take everything you're given at the start, weigh it, and produce a single answer that then has to hold. Steering happens continuously: you take what you know right now, act on it, and revise the moment something changes. Both modes count as "using information to decide," but they have opposite relationships to time. Screening spends its information the instant it produces the decision. Steering keeps spending the same information, again and again, for as long as the decision stays open.
Take the elevator lobby with the keypad instead of the up-and-down button. You punch in your destination floor before you board, and the system assigns you to a specific car on the spot. It feels like an upgrade: the system knows exactly where everyone's going instead of guessing from a hallway button press. But the assignment is also a lock. Once you're car C's rider, you're car C's rider, even if C gets stuck behind a slow stop two floors up and a nearly empty car A glides right past your lobby a minute later. The dumb version of this system, the one that only knows "someone on six wants to go up," can reassign a rider mid-wait, because it never promised anyone a specific car in the first place. Less information going in, and more room to act on what it learns going forward. The second part is worth more than the first part costs.
Fixed-bid contracting runs the same trade in a different unit. Write the full spec before a line of code exists, price the whole job against that spec, sign it. Everyone involved treats the detail up front as diligence, and in a narrow sense it is: more spec, fewer surprises about scope, on paper. What it actually buys is a decision that can't move without a change order and an argument, locked in at the exact moment both sides know the least about what building the thing will teach them. Time-and-materials work looks reckless next to that: no fixed scope, no locked price. But it's the arrangement that can act on what week six discovers in week seven. The fixed bid spent its flexibility to buy certainty it didn't need yet.
Information only has value if something can still be done with it. A forecast is worth having if you can still change your route. A diagnosis is worth having if it changes the treatment. Information gathered right before a decision freezes is worth exactly what it takes to produce that one decision, and not a cent more, because the decision, once locked, stops listening to anything learned after it.
Commitment does the damage. More information just gets offered as the reason to accept it. What's worth protecting is the ability to act on what you learn later, not any particular fact you collected on the way in. Stay uncommitted a little longer than feels responsible.
The responsible feeling is usually just commitment wearing a better outfit.