I'm happy to have experienced the free internet. Truly a jewel of humanity.
However, the good news is that we will never stop reinventing everything. The real value of the old internet was showing us what is possible.
Of equal value is that it showed us what not to do.
We have 30 years of documentation for research on exactly what a successful intra-planetary network needs to be immune to. A successful future network must build-in resistance all forms of human pyschopathology from the ground up.
Sorry I don't see how ML can help here. It seems like another thing to pin hopes of repairing an already too broken system on.
"We cannot solve our problems with the same thinking we used when we created them." -- Albert Einstein
"A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it." -- Max Planck
We are the dying generation my friend. We built it. They came. It didn't work. Surely if ML can do anything it's telling us that we need to tear down the old system completely and start again, don't you think? Adding sticking tape won't help.
edit: turning a grunt into an honest question
That's exactly the world in which the Internet grew. There were multiple segregated national and sub-national networks, and the Internet was built as a means to interconnect them. After some time, the Internet protocols ended up being used even within these networks, but that was not originally the case. And even today, there are still things like the AS (Autonomous System) concept which permeates the core of the top-level Internet routing protocols, which still reflect the Internet being a "network of networks" instead of a single unified network.
That's why I'm not too worried about the Internet fragmenting; we've seen this before. What happens next is gateways between the networks, and there are already shades of these in the VPN providers which allow one to connect as if one were located in a different network, often from a different country.
i think it already has.
the Great Firewall of China is the classic example, but I think the trend started in the west with the Right to be forgotten/right to erasure in Europe, and subsequent HTTP Status 451 Unavailable For Legal Reasons. GDPR just further cemented the split between Europe and the rest, and the new DMA & DSA regulation in the European Union finally makes it clear. The writing is of course on the wall, so countries like India or Australia aren't too far behind. Places like California also have their own "right to be forgotten", and I'm sure the US will not be left behind for too long before we see regulation further splitting their internet from the RoW. And I don't think the RoW will hold off much longer till it also splits into multiple big blocks. It's the start of the new "nationalist" internet, and I'm sure we'll all be poorer because of it.
So it can (at least eventually) function without 'regular internet', although I would still be hesitant to call it a viable infrastructure choice if the goal is to get around government control, simply from how much SpaceX have to appease the government to do anything space related.
The mesh network should be made out of common hardware in order to be viable. I'd suggest phones but those devices are owned before they've even left the factory.
"Star Wars Episode 10: The one that's not fiction."
On the other hand, if you want to demonstrate that you have anti-satellite capability it's probably a better idea to shoot down a corporate satellite than a military one. The Soviet Union shot down Korean Air Lines Flight 007 and it didn't start a war, after all.
Cryptocurrencies might be a problem in this plan, and satellite internet access itself might become a currency (since unlike cryptocurrencies, this one both has almost an intrinsic value and provides its own infrastructure that's very hard to block, where as cryptos rely on external sources of Internet access).
It also depends - drugs have consistently won the war on drugs despite being a physical product that needs a local supply chain and various anti-money-laundering and banking/finance regulations that should make it hard to fund the operation. Satellite internet access is likely to be even easier as it doesn't rely on a physical product (if we reach this stage there's going to be clandestine satellite terminals built locally, so blocking shipments of the real thing isn't going to cut it).
The only solution, apart from North Korea-levels of isolation (and even then, NK has the advantage of their population being isolated & indoctrinated since birth, something most other countries won't achieve even if they turned authoritarian overnight) would be detection followed by harsh punishment, but this has the downside of not only wasting the disclosure of detection capabilities (that are useful to the military) but also outsourcing the R&D of evading such capabilities into the open which enemies will no doubt pick up on too and use against you in a conflict.
I'm saying this as someone who once wrote a decentralized P2P mesh for instant messaging[1]. I was inspired by the HK protests going on ~2014 after hearing that they were using Bluetooth chat apps. Luckily Matrix, Telegram, Signal, etc. mostly solved the problem. Still, I don't think any amount of mesh networking would turn back the tide of Hong Kong now.
There don't need to be. You publicly gruesomely execute the first 100 or so you catch, and the practice of running a mesh node on your cell phone will fall so far out of fashion that the network breaks.
Societal shortcomings cannot be fixed via tech alone. If you can't build a society resilient to authoritarianism in the first place, tech will not help you. It can be used to increase resilience, but that's far from fixing the problem by itself.