Code as byproduct of learning
Spelunk, try something, repeat. Our work is figuring things out. Code is the byproduct.
I’ve done a few projects in PayPal’s PYUSD ecosystem. Most recently: a CLI and reference implementation for cross-chain PYUSD transfers using LayerZero.
The LayerZero docs are solid enough on the protocol, architecture, and abstractions, and they provide references for how to get the specific configuration details needed. But if you’re a builder who looking to transfer value from one chain to another right now, you find yourself patching together a ton of pieces, and you encounter rough edges: PYUSD is represented and used in different ways on various chains, some transfer paths require intermediate steps, testnets don't have existing PYUSD0 deployments. Oh, and there are libs if you poke around a bit. Stuff you only learn by exploring and doing.
Anyone developing in new spaces sees this pattern over and over. It all works in theory, from the perspective of the different major players. But how do you, as a relative outsider, come in and put it together? It's an underappreciated category of work: the end-to-end implementation that should exist but just doesn’t yet.
Someone needs to close the gap. Most of that gap is about gathering information, hitting unexpected bumps, finding the way over/past them, and unifying into knowledge. The code deliverable is the byproduct of having done the learning. Useful? Sure. But the primary value is that you showed (or validated) what's possible.