First with the Caro Kann

Haven't ever played the caro kann

4 CaroKann advance variation
6 ...nc3 Ok, new defense, but so far so good
7 ...a5 was thinking pin the pawn and attack d4, but I miscounted.
Annoying but it get sooo much worse
9. ...b5. if pxp, then bxp, b4 nxp, pxn, qxp and his queenside is open.
after nd7 and 0-0 I can ram it down the board. It's worth noting that no
one agrees with me.
10 nd2 he didn't see the skewer anyway
10 ...b4 computer wanted c4. I was thinking attack the chain at the base,
if pxb4, Nxp, and his center pawns are exposed. Further plans include bc2,
then Nd3. situation c5/d4 needs resolving, I need to castle.
13 pxc5 nc2 Ok, the thought is I force the R to b1, then Ne3 skewering the Q
and R. pxN, BxR and I've won the exchange. And the p on c5 should be available
soon.
NOTE 14... Ne3 was called a blunder by the computer, who wanted Nb4. But I
thought it traded a N and a P for a R, which I liked, especially since I thought
I'd get the pawn on c4 anyway
15...bxb1 Love it when a plan comes together
17 ...Ng6, shoulda been h6, but I wanted to castle.
18... h6 nope, Ne7 was the move and I knew it. Not playing well
30... Qa5 sigh. I'm a moron
31 Nc6 Qxa4 I wonder if RxN, PxN Rc1 attacking the doubled pawns was better. Didn't see
it in the game; too busy hating myself.

Need a bit of work on the opening, I guess.

"Need a bit of work on the opening, I guess."
This was a sideline, you cannot work out all possible sidelines. Even if you could it would not make sense as it will not occur again.
The main mistake are the pawn moves 5...c5, 8...b5, 9...b4 while your pieces still sit on the back rank.
7...Qa5 and 11...Qc7 is also a loss of time. The queen should go to b6 attacking d4 and b2, but only after the minor pieces are developed in play.
10...Qb4 was better: attacking d4 and b2

5.g3 is an unusual variation. I like your move 5...c5! attacking the centre. Qa5 was a mistake much better is Bg4(engine) although Qb6 (putting pressure on the centre also seems good). You played the middle game terribly, so I recommend working on that.

the late development of the kingside N and B is based on a video from IChess, who recommend c5 first. I'm going to rewatch it in a bit (Dungeons and Dragons first, tho) in case I got it wrong. I often delay castling with Black, but this was a bit much.

Ya'll are right, it wasn't the opening. The middle game was bad. I've just reached the stage of having memorized the openings and main variations, it might be time to study the transition into the middle game.

Actually, with 14...Ne3 you gave two pieces for a rook, not one - and that's why it's a blunder.

@sheldonleecooper515 O wow. I went over that a few times after and completely missed that. Even if I'd got the pawn back it would only have been even. Thanks!

It would be even by simple mathematics which is taught when talking to beginners: 3+3 = 5+1 = 6.
But in general, two pieces are clearly better than a rook and a pawn.
Of course, that's generalizing; there are exceptions, but not many.

You don't have to learn the kingside fianchetto (for white). That's an extremely rare setup against the CK, usually you see it only when someone is playing a King's Indian Attack, and that's rather pleasant to play against, even without any prep. Just don't castle long. :)

If you want to learn the CK, I have a short list of variations you should know: