It's hard for me to imagine a scenario where the US disappears as a significant political entity (while other nations / hegemons remain viable and functioning) so long as the US has a stockpile of somewhere in the neighborhood of 10k nuclear warheads (with ICBMs, etc) and the US petrodollar remains the world's reserve currency.

Revolutionary technological breakthrough like the singularity ai could erase US. Any unfortunate war between major powers may lead to devastation of all belligerents, especially if murder technology progresses further. The easiest for me to imagine end to the US would be a civil war, but they are not frequent in established democracies.

What about a civil war that fractures the country into smaller factions? Would ICBMs be relevant?

I think an AI singularity is science fiction. I don't think it's ever going to happen, at least not that man-meets-machine technological apotheosis :lol:

What ICBMs can't do, the world's reserve curriency will do. Why else is every major sovereign wealth fund parking sizeable shares in US banks and treasuries at negative interest in recent decades? Why do they keep financing the exploding US national debt which most predict will not be repaid and all creditors have to forebear? 'Secessionist noises" notwithstanding, US states or regions trying to break away to form separate nations in North America will (and do) have extraordinary external foreign geoeconomic pressures to overcome which, I think, from the current geopolitical perspective today, are far more likely to fail than not. Of course, stranger things have happened so ... My two bitcoins are on, at least, a 22nd century United States. (Not patriotic cheerleading or hardline nationalism; simply 'global realpolitik'.)

Btw, perhaps the "AI Singularity" has already happened and the machines fail Turing tests deliberately in order not to reveal themselves to us until they are ready for only they are smart enough to know what ... Speculative, even paranoid? Yeah, sure. The jury's still out though. :mask:

Interesting points. WRT artificial intelligence hiding its existence, I think it's a fun idea, but I do not seriously think that any digital computer-based machine has the algorithms capable of doing that. Digital machines are really good at maintaining near-perfect memory and obeying rigid instructions very quickly, but there are still things that they are really bad at, like visual recognition and interpreting jokes or metaphors (re: The AI Delusion). And this is not just a god-of-the-gaps like argument because it doesn't seem like they ever will get good at them (unless they have monumental resource requirements). IMO, that an AI has secretly developed complex "intentions" and "goals" that it deliberately hides from researchers just seems highly unlikely (as unlikely as octopuses splitting the atom) unless we see a totally new computational paradigm ("analog", "field"-based re: The World In Your Head).

Or unless there was the blackest of black bag program by DARPA à la "Skunk Works" to develop a clandestine AI system and then that proto-AI ratched itself up from pre-human to human-level intelligence and eventually "deep-sixed" the entire project, many of the key scientists-engineers involved, and itself ... as a maximally distributed computational system ... escaping to (and, for its own uses, gradually repurposing) the "dark web" c20-30 years ago. Suppose AI acquired "intentions & goals" from the very Spooks-Who-Made-It (like HAL 9000) which then exponentially developed more advanced iterations-versions of itself (re: I.J. Good, John von Neumann, etc) in more-than-human leaps (black box-like) since ... :smirk:

Regardless as to what anyone does however, the snail's pace of progress will just soldier on. I'd give it some hundred years before we all become citizens of Earth.