I don't want to get bogged down with a bunch of proactive technical learning. In general it's awesome, and I'm a big fan of it. It's really the biggest reason I pursue side projects like this one. But my natural instinct is to spend more time learning than building.
In this case I think the better thing to do is go against my usual instincts and do all my learning in a reactive way by learning just what I need to build the very next thing and nothing more.
In that vein, I've settled on the goal of building a minimal end-to-end proof-of-concept. I'll have a DB schema. A Lambda end point to manually push a new item to it, and a live site that pulls items from the DB and displays them. So it's the minimum set of things that will exercise all the major elements of the stack, DB, Lambda, Netlify host, and Elm Client. I have plenty still to learn about a few of these, but I'll only learn what I need to do to get this end-to-end PoC going rather than going through countless (and fascinating) tutorials and docs.