For ten years I asked the wrong question. What can I build? I’d look at a problem and audit my capability: what tools I had, what skills I had, how long it would take, whether I could afford the stack. The answer was always: less than I wanted.
The agent changed the question. Now I ask what should exist? And I find out whether it should by trying it the next morning.
This is what people mean when they say tools shape thinking. Not “the spreadsheet shapes how you think about budgets.” That’s true but flat. It’s the deeper version: the tools you have access to determine which questions you’re allowed to ask without embarrassment.
When the only way to ship a feature was to hire three engineers, should we build this? had a different cost than the cost of being wrong. So you asked smaller questions. What can we ship in this sprint? What’s the minimum viable scope? Where can we cut? All those questions are downstream of the constraint. The constraint shaped the question. The question shaped the work.
Now the constraint shifts. The capability question becomes a taste question.
What should exist? is a taste question. The agent doesn’t have taste. You do.
Audit the questions you’ve been asking. The ones you’ve been avoiding because of cost may be the ones that matter. The ones you’ve been asking out of habit may not.
The tools changed. Did the questions?