Once again, hope this is allowed. Feel free to move to a more braindump-y, out-of-the-way category.
• We should retain the ability to just create our own nodes (node being a fedi instance, a particular open source project, etc). But not everyone is cut out to be their own admin and that is okay. A good balance would be human scale; most people will pick a node, but some people can just create one. Where admin =/= celebrity.
• To retain our ability to pick instances, network effect needs to be reduced. Things that create network effect include: social networks, digital economy. Federation and decentralisation (and for things that can’t do that, usually public utilities become the solution) are useful here.
• One worry is instances being too big and becoming “allowlist only” such as gmail or Google Talk. To prevent this, we need a ton of nodes. Human scale comes into play there again. One good thing about admins being community members; they don’t want the server to be too big either.
• But what about things without network effects? Masto.host for example doesn’t have a network effect. Should we care about human scale there? Centralisation brings efficiencies we cannot ignore. (Related to Open brainstorming: Money, small communities and inter-community trading and communication)
Where does people-scale end? Communities? Their managed server host (e.g. masto.host?) What about their VPS providers? And what about the hardware manufactures, chip manufacturers, silicon miners and so forth? Where is the boundary?
Originally I felt the boundary would be “where network effects end” but… how does one define network effects? And people-scale comes with other benefits too (being able to talk to the creator, more humane working conditions etc that aren’t really related to network effects). But we can’t really expect smallness in things like IC manufacturing, can we?
Also, people-scale would be a more cross-species way of talking about human-scale I guess 