I played for about three hours last night. So far, I don't regret spending the $30.
SMAC probably wins on overall "atmosphere", but there's much to like in Pandora. The things quoted in Erik's post above all stand out as things I would have mentioned, myself.
The randomized tech tree works. I've started two separate games, and the tech run-up was quite different between the two. I like this, because it means there is no optimal "Golden Path" for tech advancement. It looks like the randomness is constrained somehow, to keep things from getting totally zany; perhaps the Proxy folks can comment. Also, you don't see the whole tree, all the way to end. There is a (generous) "horizon" of sorts, and as you move further down development paths, you see more possibilities open up down the way. It seems fairly well done, though I should like to finish a complete game to see how it carries through to the end.
At first, city management confused me, because I expected it to work like SMAC's, when it really works more like MOO 2's colony management. As the city population grows, each point of population get assigned to a role automatically. So you have farmers, miners, workers, and scientists. You can reassign roles at will, and lock down roles to keep those assignments fixed, if you want to. You don't place population units on city hexes directly; instead, the game seems to put each point in the best available hex ("territory", in game parlance) for its assigned role. You still get to improve territories using formers, so you can also influence city development that way. Cities start with seven hexes and expand one hex at a time, Civ V-style. The game tells you how long it will take to acquire a new hex, and lets you specify the hex that will be obtained next, if you care.
Each city tracks a number of variables, all of which can be seen on the city screen. You have: habitat, population growth, food, minerals, production, research, credits, morale, and pollution. Mousing over the values brings up a tooltip breakdown that shows how the numbers are derived. Food, minerals, credits, and research go into a global pool. I guess research was always global in Civ-style games, but having the other resources in a pool helps reduce some of the micromanagement burden. In SMAC, you'd be running crawler caravans to shuttle production from one base to another if you needed to build something expensive quickly; with a global pool, you don't need to do that so much; the minerals come out of the global pool, and the bottleneck then becomes the production rating of the city, which starts with the number of people assigned as workers, modified by buildings and productivity advances.
Instead of having discrete, unhappy drones, Pandora models happiness through the morale variable, which can be influenced through various factors like taxation rate, nearby terrain features, and I'm sure a bunch of other things I haven't discovered yet. The effects of morale act as a multiplier on resource output, with lower values penalizing output and higher values providing a bonus. It also appears to affect migration between cities, which is an interesting dynamic. I had two very crowded cities; once I established a new colony, it grew very quickly, because of a large influx of people from the other two cities. All you need is enough habitat to house them, and food to feed them all (which comes out of the global pool). Habitat is a local city variable, related to the number of hexes the city occupies, modified by building improvements and technological advances.
Pandora's production values are better than I expected. The graphics are clean and well-executed. Performance-wise, the game engine seems pretty snappy. If I had any complaints from a design or aesthetic point of view, it's that there are maybe a few too many black-and-white elements in the UI, and that the result comes off feeling a bit bland as a result... but this is perhaps a matter of taste.
My main complaint so far has to do with documentation. The game's "manual" is nothing of the sort. It contains installation instructions, and that's pretty much it; the rest of it consists of pages and pages of faction lore which tell little or nothing about the game systems. Some of this information is conveyed via tutorial tips that pop up on a regular basis, and some of it can be found in the Compendium (basically, Pandora's equivalent of the Datalinks or Civilopedia). In spite of these resources, seemingly important concepts go unexplained.
For example: There are different weapon classes (automatic weapons, missile launchers, flamethrowers, cannons, sniper weapons, anti-air weapons, beam weapons, etc.), but I don't know how or even whether these classes are significant. While tooltips and the Compendium will tell you that a particular weapon has a bonus or penalty vs. biological or mechanical targets, there's nothing to suggest why you should design units to use flamethrowers vs. automatic weapons, when both have the same tagged bonuses vs. target types. In other words, there is no description of "Missile launcher" or "Cannon" as abstract concepts in the Compendium. Should I care? Have I overlooked something?
Anyway, these are just initial impressions. Pandora is not a complete clone of SMAC, and it does try some new things. I suspect that how people react to it will depend on whether they've played SMAC before. Overall, the game seems like a solid effort, and I look forward to playing it more to see how coherent it is through the course of a complete game.