In a post-AI paradigm, we'll generate thousands of answers to problems ad nauseam. In architecture, that means creating thousands of schemes. However, when AI can rationalize all the answers, we can only rely on the one thing that remains strictly human: the gut call.
I built an architecture practice with AI from day one, which is still a rare thing to say. Most firms came to AI tools with protocols and systems already in place; I configured mine for AI instead. It can explain how the form derives from the site, how the massing answers the sun path, and how the language echoes a contextual lineage.
For most of architecture's history, reasoning was the slow part of the job. It won you the client and got you through the crit. And I'll be honest about something the profession mostly won't admit: A lot of the time, the rationalizing came afterward. You made the call with your gut and built the argument later, so the decision would look inevitable instead of instinctive.
That reasoning can still persuade a client. What it can't do is help you choose which decision to make. When the machine argues for all 10 schemes with equal conviction, its argument stops being a filter.
The act of choosing is naked. I sit with schemes on my screen and no logic to drive my decision. Something older dictates my direction: the resonance of the form and light, the proportions of curves and edges, how I "feel."
For architects and many other professions, the scarce thing is no longer analysis. It's a gut feeling you can trust, and the nerve to follow it before you can explain it. AI can make anything you ask for, but it has no stake in what deserves to exist.
