.

.

Your mom is my top 10 all time 

But seriously, none of the players commonly mentioned in GOAT lists (greatest of all time) were known for being good due to engines. Capa and Fischer didn't play in the computer age. Kasparov used engines, but is more known for his hard work. Carlsen is a modern player but is known for practical play and endgames (things you can't copy from engines).

Computers are different than engines though. Without the internet or databases it would be difficult to get good material in countries that aren't Russia for example. So without computers it's a good question whether or not Carlsen would be champ.

The more far back we go ... natural talent and novel ideas were more important. Now it's all about endless theory, endless analysis and endless preparation. And hunting for centipawns. Am I correct?

I think the old times were more novel and romantic, even though the level of play was lower.

You'll never hear "centipawns" in an interview of top players. As for endless prep, just look at Carlsen over the last 5 years (not known for his theoretical lines to say the least). Interestingly modern opening play sees white admitting there is no first move advantage (against extremely well prepared opponents), and so white is mixing in moves like 1.c4 and even 1.g3 and going for positions where there will be a lot of practical chances (something engines are oblivious to). Often by not exchanging much during the opening.

Opening prep is important but, it seems to me anyway, long gone at the days of Kasparov spending endless hours stocking up on surprise novelties in deep theoretical lines.

---

Yes, old times were much more novel. Modern pros have studied all the past players and learned from them before maturing as players themselves.

Never prepared for a single game.

centipawns are something computers have to use because they are far from perfect chess players. Admitedly they are nearer to perfect than any humans, but in a somewhat different way.

Neither have I, but our competition isn't prepared either so it's no problem.

Growing up in the 50's and 60's we never had computers but played mostly board games.  I started playing chess in the military and remember seeing an advertisement of a computer that could play chess in the 70's.  I was  infatuated by the idea of something like that.  I bought my first chess computer in the 1980's which was a Radio Shack 1650 and even though it could beat me 90% of the time the playing strength of those "toys" was really a joke until the early 90's.  I really thought that the revered Grandmasters would always reign supreme over any chess computer but after Kasparov lost to Deep Blue much of the facination and mystique of chess vanished for me. Now any patzer sitting in his parents' basement with pajamas on can say with just as much authority as the World Champion whether a move is bad or good. 

To be precise, said patzer is merely acting a relay of the views of an engine.

They can certainly say when it's a tactical blunder, but spend a while on the forums... or go to chessbomb to watch a game live, where it's clear the engine monkeys have no idea what's going on. Everyone who screams BLUNDER changes their mind 5 moves later when the engine also changes its "mind."

I still own my Saitek Mephisto Chess Challenger endorsed by Kasparov 20 years on, many hours of my life lost to it. Around 2000 strength and a strong opponent on the top setting apart from one specific random glitch where it will instantly sac a rook for a pawn on a half open h file in any position with either colour. I've read that in the early days it was very difficult to programme these things without glitches of this type, or without them making aimless moves in locked positions and flunking the endgame. Been watching some of the Komodo odds challenges on chess.comTV and its strength is incredible. How a GM/IM can be defeated with a 4 move head start or with a queen for 2 bishops etc is mind boggling, but you watch it unfold without even any obvious blunders by the human player. Amazing that we can create a machine that performs way above our own ability.  

The great game of chess will always be around. Yes people try and improve on it with about a milllion variants, but chess is still a great unsolved game.  Its a puzzle.  As soon as you solve one thing another pops up. All computers do is make the game sterile.  Mindless theory, having a machine do the work for you, blindly copying what an engine say to do.  Players proclaiming a superior position but unable to explain "why"