Hi Terry,
1) This isn't strictly true, since there are ways Villain can deviate from GTO play without giving up any EV as long as we keep playing our equilibrium strategy.  But it's reasonable intuition.
2) Yes
3) Yea, the player in position will generally have a small edge in any given hand, but assuming they both play the same number of hands in and out of position, at the same stack sizes, things will even out on average.
Regarding PokerSnowie: my answers above are all for the case of headsup play.  The whole idea of GTO play is a lot less useful/satisfying/well-defined in the 3+ player case.  The game still has Nash equilibria, but playing those strategies doesn't give you any of the guarantees about at least breaking even on avg that makes it such an attractive approach in the 2-player case.  A lot of the claims on the PokerSnowie website regarding it teaching perfect unexploitable play are almost certainly bs.  It's possible that the technical people on that project have done something useful while just the marketing people are clueless or dishonest, but I don't have high expectations.
Will

There is a brief write up here:
http://www.husng.com/content/game-theory-optimal-and-exploitative-play
 
If you have some time, and are interested in the subject suggest reading:
- The Mathematics of Poker (Chen & Ankenman)
- Expert Heads up No Limit Hold-em (Tipton)
There is also a pretty thought provoking series on approximating pre & postflop optimal play in 6max cash games @:
http://en.donkr.com/Forum/pokerarticles-at-donkr--533519
 
If you get through the Mathematics of Poker (only worth the effort if you have a pretty decent math background), you will realise that approximating GTO play for the No Limit Holdem is probably not practically possible. It does, however help you develop your thinking for street by street play which does practically help your game. 
 
A thought I had when considering this question was thinking about a simpler game:paper scissors, rock.
It is a classic game theory example for which the GTO strategy is to play randomised 33% paper, 33% scissors and 33% rock.
This strategy is unexploitable.
Does it profit vs exploitable strategies?
Take an extreme example, someone playing 100% rock.
Well, 33% you win when you play paper.
33% you lose when you play scissors.
33% you draw when you play rock.
Your EV for your GTO strategy vs this exploitable player can be calculated:
Let payoffs be:
win = +1
draw = 0
lose = -1
EV(GTO) = 0.33 * [1] + 0.33 * [0] + 0.33 * [-1] = 0
=> Expected payoff is 0 even though opponent deviated massively from GTO strategy.
 
What about if opponent is playing another exploitable strategy of 50% paper and 50% rock?
EV(P) = 0.5 * [0] + 0.5 * [+1] = 0.5
EV(S) = 0.5 * [+1] + 0.5 * [-1] = 0
EV(R) = 0.5 * [-1] + 0.5 * [0] = -0.5
=> EV(GTO) = 0.33 [0.5] * 0.33 * [0] + 0.33 * [-0.5] = 0
=> Expected payoff is 0 even though opponent deviated from GTO strategy
 
These examples can be extended algebraicly to show more generalised results.
 
GTO is UNEXPLOITABLE 
=> This means neither player can do better (more EV) by changing their strategy against someone else who is playing GTO
=> It does not neccessarily mean it is profitable by default if you play GTO if your opponent does not
 
Even if you could approximate a GTO strategy for No Limit Holdem poker (believed not possible with existing computing power) you wouldn't want to play it anyway because of the rake.
WHAT DO I MEAN BY THIS?
In equilibrium if both players were playing GTO, they will both lose because of the rake (zero edge to overcome fixed cost of playing).
In fact since expected profits from playing poker are equal to your edge over your opponents less the rake, and GTO strategy is only your most EV strategy if your opponent is also playing GTO strategy:
=> why would you even want to play GTO strategy?
=> if it was your BEST option, by logical deduction it means you're playing in games you can't beat..
 
THAT SAID, thinking about and studying game theory in poker will make you more aware of players exploitable tendencies/ranges => which you can find edges against.
For example, in the paper, scissors, rocks game being aware of the GTO strategy will allow you detect if your opponent is deviating from it and then exploit their strategy. If you observe your opponent is playing 40% rock, this conditionally affects how often he plays scissors and paper leaving big leaks in his game to exploit. 
You can apply the same thinking in poker to opponents frequencies/ranges to find weaknesses in their game to exploit.
 
In my opinion you should aim to play exploitatively, and adjust towards optimal play as your opponent makes you until such a point that you feel like you can't find an edge against them. At this point, you would not play with them unless:
intangible EV(intellectual challenge/fun + learning opportunities to find bigger edges against other players) > tangible EV(making money in games you have an exploitative edge)
 
To paraphrase Mersenneary's article (link provided above): if your goal in poker is to make as much money as possible, you should find exploitative edges and push them hard. This will not be the GTO strategy.

Pokersnowie may be the best poker marketing project I've ever seen.
Don't know if it's any good, but it doesn't really matter at this point, it's probably worth six figures even if it randomly spouted out decisions.