/tech/ - Technology

>>18861 (OP) 
Hierarchical centralization is just efficient, there is no ways around it. All solutions aiming for decentralization will face capacity overheads. Best you can do is to legally prohibit wide manufacture/distribution of proprietary hardware/software. Make it a requirement to provide full disclosure of schematics, manufacture process details and sources before entering the market, punish using rare and difficult to replicate components and techniques with extra taxes. This will discourage participation from companies and ensure small scale and quality of the network. And drop a nuke on every place in the world that would try it differently.

If you've kept track of web founders in recent times, such as Tim Berners-Lee, you'll probably be familiar with the idea of the "semantic web", the idea of which is to naturally embed machine-readable meaning into webpages.

This would naturally render irrelevant all of search engines, large websites, addresses, hosting, and specific servers, because everything would be dynamically assembled using transclusion from publicly visible sources, rather than today's opaque server-side dynamic crap or yesterday's static HTML.

This was, of course, also the original concept for how the web was supposed to work. Prior to the web, it was also how every idealized hypermedia system (Xanadu, NLS, etc.) was supposed to work.

The status quo that this didn't happen both caused and was further cemented by "web 2.0", large websites built on user contributions, effectively a second system duplicating the web inside itself in crippled form. A digital enclosure of the commons, as argued most militantly in this essay:
Why you should have a website: it's the law!
https://homepages.cwi.nl/~steven/vandf/2008.03-website.html

It has, of course, been commonly argued that, with modern netizens completely unaware of the possibility of semantic authoring, and popular tools (maliciously or otherwise) unsuited to effortlessly integrating semantic meaning into human authored data, the semantic web is today practically impossible for reasons of inertia:
http://www.incompleteideas.net/IncIdeas/BitterLesson.html

I think, somewhat ironically, AI might provide a helpful stepping stone. By ruthlessly stripmining the non-semantic web to inject it with a best guess at intended meaning, it may "prime the pump" for semantic browsing tools, perhaps incentivizing the addition of good semantic authoring features to existing tools, which would make semantically rich content the norm, gradually replacing machine-guessed semantic meaning in newer content with human-authored content made the right way, as effortless second nature.

>>18862
Centralization is the entire problem. Open sores "write-only code" is completely unhelpful in this context. What matters is standardized, stable, well documented, public APIs. Completely proprietary commercial software with good APIs is a gazillion times more open than a giant tangle of freetard spaghetti that's impossible to reimplement.

humanity ruined the internet
It's a classic example of how you can't fix a broken foundation through protocol

>>18861 (OP) 
Just replace it with Reticulum.

>>18861 (OP) 
make it work like Tor/i2p/freenet/Gnu Net by default, and make it mandatory for everyone to run a node. This way glowies or the Big Tech won't be able to control everything.

>>18871
Reticulum is pretty much that, except that if every device is a node then you will just waste bandwidth on constant announcements. If you just turned every static wi-fi router in a city into a reticulum node then you'd already have more than enough coverage for mobile devices.