My favourite prompt: advise me on the trade-offs

There is one thing I ask AI more often than anything else.

Please advise me on the trade-offs.

Which way to build something. Whether to take on a piece of work. How to word a difficult message. It goes on the end of almost everything now.

It works because of what it stops. Ask AI what you should do and it will usually find its way to the answer you were already leaning towards. You put the lean in the question without meaning to, and it is agreeable enough to go along with you. Ask for the trade-offs instead and it has to lay out both ends, including the one you were hoping to skip.

Here is a funny way to think about why that matters.

Seinfeld has a bit about the drugstore. You are standing at a wall of cold medicine, reading ingredients you cannot understand, and then you reach the real question. This one is quick acting. This one is long lasting. When do I need to feel good? Now or later?

Nobody sells you quick acting and long lasting, because there is no such thing. The clip is at the bottom of this page and it is worth the thirty seconds.

That is not really a joke about medicine. It is a joke about how things are made, and it holds a very long way down.

Engineering has a name for the most common version of it: reliability against efficiency. You can have a system that almost never fails, or one that squeezes the most out of what it has. You cannot sit at both ends, because the thing that makes it reliable is the slack it is not using.

Physics has it too, and there nobody chose it. You cannot know exactly where a particle is and exactly how fast it is going at the same time. Pin the position down and the speed blurs. That is not a limit of the equipment. It is how the universe is put together.

Then it is everywhere you look. A bridge is stronger or it is lighter. An engine makes more power or drinks less fuel. A camera opens up to let more light in and loses the depth of what stays sharp. A password is easier to remember or harder to guess. A knife holds its edge or survives being dropped.

Mine is the school pickup, and it is reliability against efficiency exactly. I leave early, early enough that my kids come out of the door and I am already standing there. What I get is being there every single time. What I pay is about an hour a week in a car park. I could leave twenty minutes before I need to and have that hour back, and some days I would not be there when the door opened. I picked a point on that line, and I know what it cost me.

The one I am wrestling with now is the business I am building, and it is speed to market against robust and durable.

Every piece of it can go two ways. I can put it inside the one app that needs it today and have it working this afternoon. Or I can build it once as a shared piece, carefully, so the next business gets it without me building it again. The second way is slower at first and faster later.

I lean hard towards durable, and I have had to be honest with myself about why. Building is comfortable. Nobody says no to you while you are building. Selling is where people say no, and some part of me would always rather be at the keyboard than in front of someone who might not buy. So the lean is not really an engineering judgement. It is a hiding place.

I am learning to move along that line, because I am never going to succeed building features in a bubble that nobody wants to pay for.

Which is the other thing the prompt does, and I did not expect it. Asked often enough, it stops showing you only the two ends. It starts showing you which end you always drift towards.

So: please advise me on the trade-offs.