Spent months assuming the hard part of building on XRPL would be the protocol stuff. Trustlines, reserves, signing flows. Nope. The hard part was learning which problems XRPL actually solves well vs. which ones I was forcing onto it because I'd already committed. XRPL is genuinely fast and cheap for value transfer. It's not great as a general compute layer and honestly pretending otherwise costs you time. I watched myself design around limitations that didn't need to exist if I'd just scoped differently from day one. The thing nobody says: starting on XRPL biases you toward shipping. Low fees, fast finality, no gas anxiety. That's real. But it also means you can build a lot of wrong things very quickly.