Ok, here we go...
I've read a bit through this thread and while I won't be voting anytime soon (lol, OU's been pretty shitty imo for a while now and I'm not gonna play it now), I've seen some pretty flawed arguments from both sides:
Pro-ban:
Unpredictability: I really, really, REALLY hate this argument as a good part of OU relies on being unpredictable in order to succeed. Yeah, it can run a number of sets and you have to sack one mon to find out what. While that is pretty fucked up and undesirable, it isn't always the case. As people have said, Aegislash can run a number of roles, but these roles need to be needed for it to be present on a team. Would a team seem incredibly weak to Hippo and Mandibuzz if Aegis wasn't SubToxic? Is Bisharp going to tear a new one through the opponent's team if Aegislash just isn't a speedy variant with Sacred Sword? Unless your opponent actually is a bad teambuilder, you should kind of think of those possibilities when mentioning arguments (like these) that are made based on high level play. I mean, you can say it has sets that screw counters to other sets and that Aegislash's offensive presence on both sides of the spectrum make it pretty ridiculous all you want, but if you can't even bother to check the opposing team's weaknesses (that may be mitigated if this one pokemon had that set) in high level play, I'm not sure for how much the unpredictability argument holds up.
Centralization: This is a very fine line in every suspect test, and this one isn't an exception. Threats are supposed to be accounted for, but to what point should penalties for not doing so be seen as okay? If a team autoloses to threat X because they lacked a direct answer to it, is it because threat X is so ridiculously strong that you NEED a counter or you are screwed? Or is it because, despite threat X having a colorful amount of viable counters, that there are simply too many threats to take in account? Will this inflation of threats with each generation's passing eventually lead to a huge amount of bans in order to have a balanced competitive metagame at all? A bit of fluff, sorry. But I have a issue with this argument too, it being the fact that EVERYTHING has some sort of degree of centralization. Especially in a case like Talonflame, who's presence requires EVERY team, Offensive and Defensive alike, to have a plan, as it's revenge killing prowess has lowered the viability of some threats, while making others completely irrelevant. This is also very bad, as now, one slot in my team NEEDS to be a Talonflame check/counter, lest I get swept due my team being part of the 99% of Offensive mons that the bird revenges without breaking a sweat. The threat of Talonflame has reduced the viability of some pokemon, if not outright removing them from relevance. We shouldn't let something that diversity limiting be here, now should we? Flaw pointing aside, I do think that Aegislash forces most mons to make room for coverage for it and makes some mons irrelevant. The unpredictability is a factor to this, as you need to counter Aegislash (or have a plan) in order not to get walled/swept by it, something that is hard to do if you don't know it's set and, thus, don't know if your counter is a counter at all. However, the issue of not knowing is already addressed above, so the only thing that remains is it's influence on other pokemon in the meta, which is in most cases merely a swap of a move that for most is not really an issue and is a sign of pokemon adapting to a top tier threat (Scizor's DPP era had more things, such Babiri berry DDtar with Fire Punch if you felt like screwing with the then #1 threat). Obviously, being in the way of some things from being usable has already been addressed as well.
Smaller things i have an issue with:
Ridiculous offenses in 150/150 is huge...just like Kyurem-B's base 170 Attack. What I'm pointing out is that OU is no stranger to extremely strong attackers and has worked it's way around these. Mixed offenses coupled with the already addressed unpredictability does make it hard to know what to switch in, but already pointed out why unpredictability is a bad argument.
There's also the King's Shield 50/50 in favor of the Aegislash user...I don't even want to point out that is a clear contradiction. These are mindgames that rely on skill (or rather, the lucky guess) and are present in many more scenarios than just this. Remember that a coinflip can go both ways (unless you're like me and get hit by 4 Zap Cannons in a row in the Battle Maison) and that the cost of screwing up for the Aegislash user is a dead (a bit ironic) sword and a world of worries less for you.
Scouting revenge killers while using a Protect esque move? That sounds like...exactly like using Protect on anything that does use it. Not really a useful thing to point out.
Anti-ban:
It counters other possibly broken threats!: Keeping something in the tier while it's possibly broken just to counter otherwise broken threats is a flawed argument, as eventually the game would be revolving around Aegislash, things that can wreck the tier without Aegislash, and Aegislash counters/trappers/lures. If threats turn out to be broken after a suspect has been banned, then they will be on the chopping block for the sake of balancing. If it's broken, it gets tossed the fuck away, because nobody likes a broken toy. Additionally, the meta might shift as other things that Aegislash pressures out of the tier might make a reemergence and turn out to counter upcoming threats. It's impossible to say that the latter will happen for certain, however, so don't quote me on that.
Making comparisons to other metagames: I'm not sure how this works unless you actually played the meta you are referencing to. And even so, it should only be a part of your argument, not the entirety of it, as meta's can change wildly over the ages (just look from DPP to BW), so situations are never the same in one meta as in the other.
Sorry for a lack of Anti-ban flaws, but I'm tired right now and can't continue to read the thread for more flawed arguments. My two cents.