Yeah… as fun as the politics of this are fun to discuss, the implementation realities seem to be entirely overlooked. Also the economic realities.
MLPchan currently operates on an EC2 instance behind a Cloudflare business plan. The former costs somewhere between $50–100 and the latter $200 per month. MLPchan was previously and I believe plans to again be on a dedicated server, which ran/runs a more flat cost in the range of $70 a month. Anonthony pays most of this out of pocket and the rest with donations (which last time he posted figures was paying for around half of it, I think).
Ponychan, by contrast, is running on six (last number quoted by Mithent) virtual hosts which cost $5 a month each. In other words, it costs ten times less to operate Ponychan than it does MLPchan.
Whoever has the keys to the kingdom ultimately is whoever is paying the server fees. So if Anonthony is not the admin, as seems to be the plan here, who's going to pay the bill? I don't know if Anonthony is really going to pay upwards of $300 a month (minus donations) to keep a site running that he doesn't even own.
That's economic. Now technical,
Tinyboard vs. Kusaba has been brought up a lot. Migration is mostly an issue of time more than difficulty. Ignoring those, in terms of features:
+ More aesthetic design (in-place editing, dialog box for reports, cleaner hide boxes)
+ More scalable settings page (not a dialog) 
+ URL tags
+ Stronger spam detection
+ Mature filter
+ Auto-refresh, etc. turned on by default
+ Better handling of >>citations
+ More functional watcher
– Timestamp for edited posts is ugly and obtrusive (sorry but this really bugs me! On Ponychan it even adds a <br> under it for some reason…)
For what my opinion is worth, having worked on Tinyboard it's a fairly robust system (as robust as you can make PHP anyway) and not at all a problem to work on. There's a few questionable things I stumble upon while coding on it, but nothing that's on the scale that prevents me from getting what I want done.
I haven't worked on Kusaba's internals at all to know what it's like on a deeper level, but I know a number of people have complained about it in the past while working on it.
Having a look at the feature list, Ponychan's appeals are more aesthetic than functional. These features aren't so much dependent on the underlying architecture as they are on the programmer having a keen eye for design and applying that to the user-interfacing code. These things could more easily be ported over to MLPchan than could the more functional things behind MLPchan be ported over to Ponychan.
From this, being only what I know, if a merger were ever to occur, I think the decision is pretty well in favour of using MLPchan's current underlying architecture. The question is then upon whom does the responsibility lie to port Ponychan's features over to MLPchan's code base?
On the topic of names: I'm not a PR guy, but ponyboards.net sounds like the least bad solution.
 Changing to EC2 resulted in a slower server per dollar but provided more flexibility against DDoS attacks. I believe with Cloudflare business plan this is now somewhat unnecessary.
 That said, PonychanX settings are separate from the other settings and handled in a fixed-position dialog that overflows on my screen size (1024x600), hiding both the save and cancel buttons.
 In a change six-months in the making, MLPchan should have this as well as a few other tags and a suitably extensible system sometimes soon. The main reason for the delay is that Macil and I are somewhat anal that it work for every edge case.
 On Ponychan, cites to posts not on the current page don't show on hover, nor do they expand on click.
 This is what I've been told, anyway.
 Tinyboard (and hence MLPchan) uses Twig for templating, which in my opinion is one of the best template languages that currently exists altogether. By contrast, I'm fairly sure Kusaba templating is done in pure PHP.