Step-by-step interactive guide from idea to output. Design an ICM workflow alongside Maya, one decision at a time, and watch a messy process become something an AI can run again and again. Right now, Maya has to try to remember what needs to be done and where each piece of information is kept. She needs a better solution. https://claude.ai/public/artifacts/7c9618f4-8324-4048-902c-cbcf77c9c102 Let’s build one with her. 😀

I spent this week doing something I don't see talked about much in the vibe-coding world. Instead of just shipping my site, I tried to break it. I built it careful. Rate limiter, guards, the usual. But reading your own code never catches anything, you wrote it, you're blind to it. So I pointed an AI at it and told it to act like an attacker. It got in in about 40 seconds, through a header I forgot you could fake. Then it found two more holes on the poll, same root cause: a security switch flipped on, but the rule behind it let everyone through. Dashboard said protected. It wasn't. Fixed it the same day, then ran the attack again live on the patched version to check the work. What's stuck with me since: this whole space is optimized for building fast. Almost nobody talks about building it so it lasts, or so it doesn't quietly leak everyone's data because you never thought to ask it not to. That maintainable, secure side is the part I actually care about, and it feels underserved. So I'm curious what you all are seeing out there. Is there real appetite forming for the "build it right, not just fast" angle? Or is the market still mostly heads-down on speed and model comparisons? Trying to read whether this is a lane worth going deeper on. Video where I walk through the whole thing: https://youtu.be/mdQ8xiKWTc0