Showing is cheap, knowing isn't: What changes when the cost of prototyping collapses?

Certainty used to be the price of admission. To earn permission to move forward, we produced the evidence of confidence: personas, journey maps, service blueprints, weeks of research artefacts, and a prototype built to prove the direction. That cost has now collapsed. With an AI partner I can prototype a solution in an afternoon. We can move forward holding a few open assumptions and a few honest gaps, because the prototype will tell us whether the direction is right. Certainty is no longer what you pay to get in. It's what the next step produces. That sounds like pure upside, and mostly it is. But cheaper steps don't make the hard questions easier. Showing an idea is cheap now. Knowing whether it's the right one, knowing what to build and whether you should, is exactly as hard as it ever was, and arguably harder, because the prototype is so convincing. In this talk I'll show you how to build a working prototype live on stage, and show you the moment that AI changed (not how FAST I worked) but HOW I worked. I'll be honest about what genuinely surprised me, and share with you the three biggest things I'm still wrestling with. I'd rather show you what I haven't solved than sell you a tidy story. If you're building software in 2026 and quietly unsure whether the ground under you is solid, this talk is for you.