Lady and gentlemen, I'm here to break up the lovefest
Have you actually read this forum? 
(1 of 4) "Deals that have been paid in full cannot be cancelled" : really? Have you ever tried telling Darth Vader that one of his deals cannot be cancelled? Right.
I wouldn't consider this a bad design decision tbh; more a way to defend the AI from unscrupulous play. There's better solutions (penalties with all races), but it's not as bad as letting the player get away with picking up a huge load of techs on a buuy-now-pay-later deal and then allowing him to walk away from it the following turn.
2) Un-abandonable colonies : coupled with a mandatory large empire penalty, which can only be cancelled by using up a race feature ... come ON dudes!
I'm going to take a guess here and say the purpose of "space tourism" is to offset the garbage planets that take approval away from your quality worlds. Well, 3 extra bc per class 4 mudhole that another race colonzed and that promptly rebelled over to me is poor compensation. I think that tourist structures really should have a negative maintenance cost, ie a base income. Does it make sense to anyone that a "port of call" would continually lose money?
Destroy planet. I'm afraid this one is a Learn to play issue on your part.
3) Divided tech selections: I cannot rage enough at this left-hand, right-hand, middle-hand nonsense! Did they think that turning the tree concept on its head is an innovation? Alright as a novelty it worked, but after a while I get to thinking "so why can't I research both things, or all three?" In fact I would have preferred that picking additional techs in a set cost less discovery points (you are already bogging down, instead of spending your research on higher tier sciences) or if owning several pieces of a set would make a subsequent tech on that branch cost less. . It Makes sense, it's called synergy and it exists everywhere in nature! Anyway why are these unchoosable techs appearing in groups of 3 always. In reality that would be statistically very improbable! (yes I know there are a couple sets that only contained two). It seems like an alright mechanic so I can take it or leave it; however there's one branch I can't stand and that's the life support. It's inconceivable that you have the entire useless branch of life support, and in addition a choice under propulsion to increase range by --wait for it...-- TWO tiles. Just makes me really worry for the eggheads of the future, you know?
Not really seeing this as a problem either, tbh. Specialization gives the player an interesting choice to make. There's some serious issues with them being hilariously unbalanced (there's almost always one 'right' choice and two wrong ones), but the actual concept is OK and gives the player a good reason to engage in diplomacy - and, indeed, most players who turn tech trading off don't bother to talk to the AI at all.
4) Player negotiation: no, no, no! Everything wrong! Race reports should be improved and fully available during a chat. "Hi, custom race guy, I have NO idea what your strengths and penalties are, but you already want to trade for planetary invasion..." When conversing with an AI, what are their techs? Since I'm already able to see what all their planets are building, what garrisons they have, the minute details of the makeup of the garrison, etc etc, should I not be able to calculate what discoveries they've made, more or less?
Um, you can see their techs during negotiations. You can see literally everything they have, in fact. If there's no techs listed, that's because you're comfortably out-teched them. Having reports available on the actual convo screen would be nice, though; as would being able to see your relations within that screen.
The biggest outright design screw-up, by far, is LEP, which just doesn't work as it stands. The design for this is fairly poor, and any way of making it work is an equally poor fudge. It needs an outright, ground-up rethink and to be either shifted to something other than approval, or for approval to be heavily reworked. Other stuff mostly just boils down to bad balance implementations - output bonuses from buildings being too high, maintenance and mid-to-late game construction costs being laughably low, carriers being just plain ridiculous, the AI's valuations in diplomacy being completely insane. Diplo as a whole still feels a lot like a placeholder, in fact, and I'm hoping we'll see a fairly large patch that reworks it.
But on the flip side, there's some very good fundamentals buried under these problems. The pop=production concept is a good move when it's balanced right (it presently isn't for the most part). The way ship combat works is also a nice idea (even though it too could use some balance tweaking) and actually makes fleet makeup reasonably important, though it needs more in-game signposting to make the player aware of how roles work.