The software on my desktop is 100% open source, but most of my time online is spent using proprietary code. To me, that's things moving in exactly the wrong direction. And I think I've figured out the reason.
Here's what's wrong: an application, as soon as it has an online aspect, transforms from a product into a service. Building software is fun, or at least rewarding; keeping a service running day in, day out is much less so.
Worrying about DDOSes, servers falling over, users forgetting their passwords, security updates, keeping up with privacy regulations - it's literally thankless work, requiring none of the creativity and discovery that writing new code does.
This is the appeal of "serverless": the developer keeps the fulfilling parts, and passes the drudgery onto someone else. But serverless is a bad fit for the open source world. Asking an open source developer to give their code to the world - that by itself is a lot to ask. But then getting charged a little bit of money every time their app gets used - that's kind of a raw deal, especially if the app becomes popular.
Edsu makes writing an online app - multi-device, multi-user - more like writing for the desktop: you just provide the code, the user brings their own resources - notably storage, in the form of what's essentially an online hard drive. It gives control of data back to the user, and it frees the developer from the hassle of maintaining servers or the cost of paying for serverless.
This small move ends up changing the landscape in surprising ways - from changing expectations of trust, to making forking work again. There's a lot to talk about - I'm looking forward to writing blog posts about these things, along with various aspects of using and developing with Edsu.