there is a strong temptation to stop
thinking and just let the machine do every-
thing; time being limited, the result can be
a shallow anno-Fritzation

You raise a point that others have raised and it is a good one. I
have read also in a number of places that now the trend is toward
concrete analysis rather than understanding the position, with
computers having pushed us to this. But in addition to what you
mention --- the trend to just let the computer do it all and then to
blindly believe it, this trend accelerated and supported by the
amazing things engines like Rybka can do --- we tend to forget that
chess for fun (even very serious fun) is played by people, over the
board, and the kind of concrete analysis featured in Chess Lice and
many new books is all but impossible for even the world's top
players. When real people play real chess, positional understanding
and general principles still play a large role, as they must.

A good case in point: this past weekend I was playing over a game from
the marvelous book of the renowned Zurich 1953 tournament. A position
was reached (I think in the third game in the book; I don't have it in
front of me) in which the annotator (Bronstein) though Black had a bit
of an advantage, and mentioned a move for White (one that I would have
likely played) as being unappealing. I didn't see quite why so I set
up the position with Rybka and poked around for 20 minutes or so.
Interestingly, though Bronstein thought Black had an edge, Rybka gave
White anything from a 10 to 25 centipawn edge depending on the
continuation, this based on deep and complex analysis with some very
unobvious moves in the projected lines of play (the move I liked and
Bronstein did not like was indeed inferior for White, but only in that
White would have less of an advantage).

But in a practical sense what does this really mean? A 10 centipawn
edge is dubious at best for a human player. 25 centipawns is better
but still it's something that a computer understands and a human would
have a much harder time converting into something tangible. And, go a
few plies deeper and we swing another 10 cp in either direction. We
are fooling ourselves if we think this is any sort of practical guide
in real-life play.

In fact, as the position stood, and as it would be played by humans,
even top grandmasters, Black did have an edge simply in the fact that
there were positional considerations that humans could understand and
deal with. Those considerations would have been overturned by
extremely deep and complex concrete analysis of which only a computer
is capable. The computer would have perhaps won with White (certainly
not have lost), but in the real game, Black pushed his "real life"
advantage and made a win out of it.

To take this exposition to an unexpected conclusion, this all explains
why I like Sanny and I like GitClub. They make very human errors, the
kind of errors and blunders that I make. I can relate to them. I
cannot relate to Rybka on any level.