Who cares? The game started at 1991, and all 28 years war AI was equally terrible. To say it was better in previous versions would be direct lie, I'd think. It never was. So, why it would become suddenly netter now?
I don't think that's really a fair argument. It's like saying "well, politicians have been corrupt for centuries. Corruption is bad, sure, but it's been happening for so long, so why should we care about it?". There are people on this forum who think that the way the devs have always approached AI is not the best way, and that it's time to make a change to it.
I can't keep up on Deity difficulty. Or even Immortal. But, if it were up to me, we'd have the AI's warring be improved considerably, and the difficulty setting would have the AI's strategy/tactics scale. The highest difficulty would have the most advanced tactical AI, and would probably be more intelligent in other manners as well (prioritizing Campuses/Theatres over holy sites most of the time, better district placement, etc.). That way, we wouldn't really have to just give the civs a ton of extra settlers at the start, which makes the early game super unbalanced (for a lower-difficulty player like me - the extra settlers for AI are the reason why I don't play the highest difficulties), and require the human player just play the catch-up game and have an easy time later on in the game because the AI can't stay powerful.
In an ideal world, the tactics would also vary based on the leader, and maybe even the match itself, to keep things interesting - it might not be as fun if all the AI leaders just play the exact same way all the time. Agendas should still be taken into consideration. Example: maybe in some games, John Curtin won't have any problems being a jerk because he knows that if you declare war on him, he'll get doubled production. Or, in other games, he plays the way his agenda probably intends, where he'll fight against a warmonger to liberate cities specifically so that he can get the production bonus for longer, but otherwise won't be aggressive, and only fight if you initiate it (again, bonus production). In any case, though, he'd prefer to settle cities in positions where he can reach high-appeal tiles for the district yield bonuses, and also try to map out district placement to maximize appeal bonuses AND adjacency bonuses. He'd plant cities on coastal tiles, but only beside the highest appeal tiles and not right on top of them, or else he won't be able to benefit from the yield bonuses. Does John Curtin think like this in games already? I don't know. I don't think he does. But maybe he should.
So yea, this discussion was about AI's combat skills specifically, but I think that if the devs improve the AI in that regard (which they should!) then they should consider taking a look at AI decision-making overall. After all, it often is connected. How are the difficulties going to scale if the AI's combat skills are improved? Would they be improved at all skill levels, so the lowest difficulty AI fights just as effectively as the highest difficulty AI? Is that really fair? Maybe the lower-difficulty AI shouldn't be
stupid, but at least shouldn't be a strategic genius like a Deity AI might be. And while we're at it, maybe we can improve the AI's non-combat related decision making in higher difficulties, and then we can remove the whole "30 extra free settlers at turn 1!" thing.