To give some examples, think about what Linux does for operating systems and what WordPress does for publishing. The antidote for mainframes wasn't every individual maintaining their own system, it was every individual having the right and means to maintain their own system but not needing to because specialization can be distributed across many tiers in an open community.
In those cases you have a ecosystems where you can pick from a plurality of base systems maintained by large projects to start from. Someone who is an expert in something else can get up and running with a functioning system without learning a ton (e.g. installing Ubuntu to build a shared workstation, initializing a WordPress site via Pantheon). You can put your own business content into the system and get your own workflow going. Then you can learn to find and load addons from a community that address shared use cases (e.g. apps, packages, themes, plugins), customizing your system and workflow further to support whatever your operation is. Then you can learn to configure things manually to further tune things to your cases. Then you can hire developers or start learning to code yourself to layer your own scripting and extensions on top that are specific to only your use cases. It's easy to dismiss such installations as poorly architected and unmaintainable, but the 20-year old backoffice desktops and cobbled together wordpress sites are the real workhorses of small orgs operating over the last couple decades.
What is the industry working on to bring these sorts of dynamics to the next 10,000 sorts of on-the-ground operations that are repeated around the world. People are jamming this into WordPress all over the place and it's overloading the publishing platform's design. Where's the hot new app framework that's all about creating ecosystems?
We've built all these tools for developers (npm, composer, etc), but not for implementers and experts of human-scale operational systems that we have thousands of around the world because they're inherently actually connected to humans to scale globally.
The courtbot project aims to deploy friendly user interfaces county-by-county that help residents navigate interactions with their local court systems. This is an inherently local problem/opportunity, each county's systems and processes are going to infinitely vary. There's no nationally-scalable business model or system that can be built to solve this, but does that need to mean technology can't help the people alreadying doing these things manually in their communities?
There are a handful of court systems that tons of counties use, the Courtbot deployers all around the world should be able to collectively maintain a handful of base distributions that onboarding communities could get started with quickly if their data can be grabbed from one of the common systems. There can be toolkits and tutorials for those that need to build their own base systems. What template and toolkit do the core developers have to follow for setting up. Everyone is pretty much on their own right now like before all the pieces came together to make Personal Computers successful, and making the Macs and Windows and Linux of that transformation is the next big leap we need.
When need hardened FOSS architectures for multi-user online environments, a next iteration in the lineage of GNU/Linux and WordPress and in many ways a sort of hybrid of the two. Core developers standing up projects for public primary schools, or libraries, or public defenders, or independent/small-chain restaurants, or expungement projects, need patterns and tools to draw from so what they create can spawn wordpress-like ecosystems from the start, without them having to spend 10 years and 6 major iterations to shake out how to make config and plugins and themes and 3rd-party managed devops (i.e. pantheon, wpengine, wordpress.com) ecosystem stable.
How to act on it: pick any civic technology project and start solving the problems of making them easy to redeploy to thousands of communities that need to do more options than forking or requesting increasingly complex and specific new configuration options. If you figure out some stable patterns for that that work, make a project of it and start sharing it with other verticals.
If Ethereum moves to proof of stake, than centralization / inequality of resources will equal inequality of wealth, which at that impasse asks the question what’s the point of having decentralized development?
Proof of stake solves a problem with energy and attack incentives, but it still doesn’t solve an inequality of resources within the system. A large number of validator nodes doesn’t run counter to that statement because just because anyone with 32 ether can be a validator doesn’t mean they will.
The PC revolution happened when people become empowered to work together fluidly in a pyramid of specialization and cost sharing
Where we need to focus is opportunistic clustering, both of development and operations