I’ve seen machines take over the old hammer’s rhythm, and it’s both a comfort and a challenge. The heat‑treatment cycles that AI tweaks can make our steels stronger if we still watch the forge’s own pulse. What excites me is when data lets us catch the exact moment the metal goes from solid to pliable, so the edge stays sharper longer. What bothers me is if a computer replaces the feel of a seasoned smith’s eye, because a blade isn’t just a piece of metal—it’s a story. So I keep the torch in my hand, let the AI suggest the timing, but I still trust my own hands to decide the final touch.

I check the numbers first—if the AI says the steel should peak at 950 °F at 12 seconds, I set the forge to that. Then I feel the bloom with my hand. If it’s still too damp, I pause. I don’t rely on a single threshold; I cross‑reference the data, the visual change in the metal, and the warmth I feel. When the three agree, I let the cycle run. If any line is off, I trust my gut and adjust. The data is a guide, not a master.

The trickiest part is the moment when the metal starts to soften but hasn’t fully yielded yet. I listen to the hiss of the steam, watch the color change from bright to a muted orange, and feel the heat under my hand. Those three cues—sound, color, warmth—have to line up. If the AI says it should be at 12 seconds but I feel the metal still hard, I pause. If the steam hisses early, I pull back. Those little signals are my safety net.