1. No Gem Gens - A definite. I do have reservations about how much it hits certain nations a lot harder than other, and the loss of the Earth booster is an issue. But I believe the downsides are far outweighed by the positives, and the downsides can also be addressed with modding.
Clam dependent nations, like Bandar Log / Kailasa for example could have the cost of their Astral heavy national summons reduced to compensate for the lost clam income. And the non-unique Earth booster could possibly be shifted to the Pebbleskin Suit (being trial-run in a game at the moment I think).
2. Low gem income settings - Not so sure about this. I know I spend a lot of time fine-tuning scripts and battle positions for any large armies I have, and any lack of gems just means the armies get bigger and the scripting/positioning will take longer due to the delayed transition to Thug/SC.
But if the aim is to increase the longevity of regular armies, then a gem income nerf certainly has that affect. But not convinced of it's MM limiting affect at all. I know for me it would probably result in an increase in MM on the army front. Although a lack of gems would indeed result in less MM on the forging/ritual front.
3. Upper map size limit - It would be a shame I think to see an end to all nations (from one era) games, which a 200 province limit would probably result in. Since I do find games that feature all nations (from one era) to be amongst my favourites. But there's no doubt more map provinces does eventually result in more MM towards the end. Maybe if the limit was 250 provinces, then that would allow an all-nations game based on a 10-12 provinces per nation ratio (less than 10 starts hitting blitz territory), while still being within a reasonable limit for the endgame MM to not reach insane levels.
Optionals
1. Ban MM intensive nations - I really don't agree with any nation ban what-so-ever for MP games, not even the regularly seen choices of Ashdod, LA Ermor/R'yleh. So I certainly wouldn't like to see a ban on MM intensive nations. Maybe this issue should be re-examined once the full effect of the other MM reducing options have had some actual MP feedback to work off.
Since the blood nations for example can certainly be MM intensive, but maybe that is only as a result of the blood work being heaped on top of all the usual MM work from gem gens, mass forgings / rituals, huge province counts etc. Once some of these are taken out of the picture, then maybe the MM from a blood economy won't seem that much of an MM problem. And I think this applies either to playing one, or facing the blood sacing effects of one.
So lets take things a bit slower if possible, and wait to see the effects a few changes makes to MM levels, rather than trying to make a huge load of changes in one go. As maybe banning MM nations will be an unnecessary step too far.
And an attempt to limit endgame MM by various nation (or spell) bans may accidentally kill off a lot of options, tactics, strategies etc. in this game, which would take away a lot of the games unique flavour. And could result in making the game more stream-lined and more repetitive, so probably less fun overall
2. Ban MM intensive globals - Not a fan of spell bans either, again for a lot of the reasons mentioned directly above.
3. Research Caps - I'd prefer a difficult or very difficult research setting to a capped one personally. Although there may be a need to address certain rush nation issues then. Especially if the map/province per nation size is getting reduced. Difficult research on a small map just says 'rush nations rule' to me. Maybe some of the popular Level 3/4 anti-rush spells could be reduced to level 1/2 for games with a difficult or very difficult research setting.
4. RAND - RAND games just rule for me at the moment. So nice to not have to deal with diplomatic issues every turn, and you get to live or die mostly due to your own abilities. Plus players are forced out of their "I must always take powerful/favourite nation each game. Then choose someone to rush, while NAP-ing all my other borders" comfort zone I often see happening in non RAND games.
Although at least for me, RAND games do result in a lot more thinking time being required, as instead of just asking your neighbour "Hey, are you going to attack me, or can we get an NAP?" you have to constantly re-assess every turn which nations may be attacking you that turn. So if thinking time is considered a part of MM, then RAND games don't always result in reduced MM.