No need to imitate giants, meaning there is no reason to use k*s. We have declarative systems like NixOS and Guix System, there is no need to use paravirtualization, who is born for commercial reasons: selling ready made stuff to those who do not know at a price you can't sell for a snippet of nix/scheme.
A company can simply develop their own decentralized infra, hosted by their employees. If it will need more it will also have money to buy sheds with p.v. and storage with good enough fiber links to have their own small datacenters.
I was thinking more of a cooperative operation. So people could totally join and quit on a whim and I think k8s would help with automatically moving your app were there is available resources.
I do not say it's not possible, just from history every "community of good citizens" never last long nor perform well. Beside the fact containers are a very absurd model as full stack virtualization before them.
If you really try to draw a big picture, we actually waste immense computing resources just to commercial evolution needs, to go "autonomous" we need to avoid copying the bad commercial model but introduce a better one. Declarative systems are MUCH better, with much less attack surface, easy development, documented evolution, ... they are not "a FLOSS version of $giantname solution", they are another paradigm.
Similarly we should avoid working on VDIs which is the giants way (let's say Amazon Workspaces, Windows 365 etc) so not going let's say Guacamole but going on bare desktop computing syncing just files, using logistic to send iron, configuration management to manage it from remote etc. We should avoid NextCloud/other clones of giants webapps to works instead locally with different tools, let's say do use R/Python not a spreadsheet. Use email with a modern MUA (we currently miss, for non techie) not a web app for development and so on. Not trying to mimic someone else paradigm but proposing a new one better fit for our purpose.
Let's say I want to setup some blog, portfolio or fun experimental site but my internet connection is not perfect. I'm not alone on this kind of problems: we should be able to pool our resources and have those websites be delivered by whoever has a working connection to the computer requiring access.
I think it could be done with k8s but maybe not. And websites are the easy model as they don't require a permanent connection between "server" and client, with things like IRC it should start being fun.
You contact $WeHostWebsites a company that do what the name imply, they have no datacenter, instead they are 10 people offering the service, some ready-made static website generators with a portfolio of themes, option for ecommerce etc, hosting their infra in their own homes. You pay them few $currency per year and you have your hosted website. At a certain point they need more iron but not much more people the owners buy a shed aside his/her own home and add the need iron, another own do the same...
A day their business goes bad, they still own the iron, they can sell if as a service for others who need computing power for a short period of time or they can reinvent themselves in something else because they still have something tangible behind them, not a rent of someone else resources. That's the model.
A company can simply develop their own decentralized infra, hosted by their employees. If it will need more it will also have money to buy sheds with p.v. and storage with good enough fiber links to have their own small datacenters.