@Voremonger
i don't disagree with your reasoning, but am biased towards a player-choice fix to this problem, rather than something applied automatically. I suppose this implies AI decision making issues, and may not be ideal from that perspective.
re buffing capital only.. might be a reasonable fix, but doesn't it essentially reproduce the same problem you mentioned re: early defensive building? given that in this case it will then be restricted to capital, again it might be preferable to a more standard building -- but there's no choice for the player; weighing risks/benefits is what civ is all about. Consider as well that human will be all the more immune to these early attacks, too (okay, they're rarely seen from AI... but maybe thats the problem)
re 2-plot range for cities from start, this is possibly the most effective defense against the archer rush, but removes the progression of building up your city's defences, and maybe makes early archer conquest altogether impossible?
re 1-range archers.. this strikes me as a pretty significant archer nerf, and game changer to early tactics, but i'd have to play it to see.. imagine it would need other values tuned as well
I suppose my concerns could be addressed by just tuning w/e solution is applied to the right values... make early rush harder and rarer but not impossible and extinct.
As human, when i start near other human, i make sure to bring my pathfinder back early to watch what's being built and adjust my local strategy accordingly.. early rush is only a problem if i have too many neighbours or some other unique situation.. its probably asking too much that AI be made really competitive doing the same, as nice as this would be... skilled human is also not as prone to having units plinked off in weird terrain configurations as AI, which is probably the crux of the issue for AI using army to defend from this.
brainstorming further here,
what if warrior were given some kind of city defence promo? iirc, existing promo structure can have a unit buff adjacent city defense to some %. not altogether sure if the same exists for garrison itself but it strikes me as easily implemented, and probably more desirable than the adjacent bonus. Cities could thus get a defence bonus through warrior garrison. Could have it disappear on upgrade, or maybe have it stick but not available on subsequent builds, to add incentive to building warrior over archer in ancient, and keeping the former alive.
Anecdotally, I'm in a deity game on huge marathon w/ 43 civs, playing 12-1 VP since early december. I didn't go for full 4000 bc rush, but as I added up the terrain and my neighbours, I switched to it just in time to catch most of the nearby AI before or as they got their 2nd city out.. wasn't quite as quick maybe as the rush you described, but just as easy... took out 3 very close AI rivals, 5 cities, by mid-ancient. By the time I had the first wrapped up, the 2nd had archers, but they were dispatched all the same, and the 3rd as well. Now, the terrain was impossible for defender on this map, even for the amazing VP AI, and would've been very difficult for human too (massive hilly forest playing iroquois).
An extreme case maybe, but probably one for which a defence should exist. I doubt most of the suggestions in this thread would've allowed for a full defence in this niche situation; just delayed the fall of each. The one thing that might've, though, is if the neighbouring AI immediately war dec'd on me as it became clear I had advantage in my first war. Either that or if one of the AI rushed me themselves, before my rush even got going. They each started hating on me for war mongering pretty quickly, but didn't actually do anything outside of denouncements until it was way too late. If archer rush is a viable strategy and nothing else is done, AI should select it appropriately often, too. Can anything be done there?