Caro-Khan is the most Strong Defence?

This makes absolutely no sense.

Where can I see these computer games, or at least the results?

https://tcec-chess.com/

Thank you

And therein lies the problem with it, that it is hard to press for a win, statistically, if you want to win tournaments. That said I was amused by the Rest of the World versus Karpov game where he made winning with the Karo Cant look easy.

I see a lot of potential in the positions you could get from the CK. Black can always get an interesting game from it.

In other openings, white can choose to play sharper lines or exchange lines.  In the future, I am taking a fairly uneducated guess that many of these will be potentially drawn or slightly better for white. 

In the CK my "hunch" is that with blacks flexible pawns and good pieces, we will see that most white attacks will be able to be defended against and white would move to a more positional approach to avoid inferior endgames.

Again, its all speculation... But how hard would Fischer have laughed if you told him the Berlin is the way of the future.

Do you get more Panovs than Slavs?

That line doesnt show the Classical variation of Caro- Kann line. 5..... Bd7 is very doubious line of Caro-Kann and Jeoren intentionally put doubious line in TCEC to get some exciting play for win or loss. (As two engines with classical opening, draw is 96%)

Better variation is .   5....... Be4, which is suggested by both Stockfish and Lco.. That line is extremely difficult to play for black and only top 2 engines can play for draw as black. Other top 5 engines like Komodo or Ethereal will lose . 

Yeah, I don't know about 5...Bd7 being "very dubious". If it truly were that bad, Lc0 should not have been able to hold a draw, no?

5...Be4 is definitely a tricky line. I don't like it, as a CK player, due to the difficulties that you mentioned.

4...h5 (instead of 4...h6) is also easier for Black to play (at least at the human level). But, like you said, the chosen line made for an interesting game!

Horatio caro believed that 1...c6 was so strong that white shouldn’t play 1.e4

Unfortunately nobody gave him two farts.

Firouzja begs to disagree.

1. e4? ...c6!

and now White hopelessly lost, after such a terrible blunder. 

(Kidding, of course.)

I don't really like the caro-kann because of how it limits knight development.

This user is not accepting friend requests 

Holding a doubious line requires playing extremely accurate moves. It was draw because it was on very long time control on extreme hardwares. Even with that TC , I bet black will not get 96% draw like other classical lines.

In case you put the same engines in blitz/ bullet, black chances of losing will be much higher.

If you look at book exit, Stockfish says, +0.74 that means the line is doubious. 

I will repeat again, if you let Stockfish or Leela run for 1 min per position, they will never play such doubious line like Bd7 and let opponent have high chance of winning like 0.74.

If you dont believe me, run Stockfish/ Lc0 yourself. 

I suppose I don't put that much faith in engine eval numbers (yet). They still fluctuate (and sometimes prove to be wrong) as the game progresses.

Yes, engines are the gods of chess, at this point. They see more than humans ever will. But they aren't yet infallible. Sometimes they're flat-out wrong.

I'm sure that, eventually, engines will be perfect and their eval scores will never be wrong. But I don't believe we're there yet.

Case in point: this was from game 75, move 56:

Stockfish thought that this position was completely won for White (+2.49!). And it gave White a 93.6% win percentage.

The position, however, was completely drawn.

So ... even a titan like Stockfish still gets it completely wrong sometimes, proving that you can't always trust its evaluation score.

I do agree with you, though, that the line played in that game wasn't the strongest for both sides. Though I'd say that the less optimal line really began at 4...h6.