I found 13.Bd3 is more commonly played at master level, such as in this game:
@1
"Is the theory presented in Beating the Sicilian 3 still good or are parts of it out of date?"
++ Yes, of course. Any opening theory book is obsolete while being printed.
"no longer the best resource for White against the Sicilian, then what is?"
++ A data base.
A message to the OP.
Take what the bit calls a mistake with a grain of salt. Too many younger and lower rated players are relying solely on artificial intelligence and that is a mistake. Here is why:
1) Human GMs beat artificial intelligence often.
2) Artificial intelligence is phenomenal at calculating forced sequences, even ones that are 25 moves deep that force the win of a critical pawn. BUT... They overvalue material above king safety or initiative. If White is down a pawn or 2 with the initiative, but it does not see a forcing sequence for White to either mate or win the material back, it will claim advantage Black, despite half his pieces taking a dozen moves just to get out.
3) Chess.com uses Stockfish 11. 15 is already out. In fact, I believe it is at 15.2.
Computers are great for tactics and 7 piece or fewer table bases. Otherwise, they are quite weak. You want to beat a computer? Avoid wild, tactical lines. Get into slow, positional stuff and you'll be playing nothing stronger than MAYBE IM strength!
In chess.com's Han Niemann report they stated that if Magnus Carlsen played Stockfish 100 times the expected result would be 0 wins for the human.
This is not disagreeing with the assertion that engines have flaws, just that I don't understand how Stockfish could be rated almost 4000 strength if GMs can beat it. I only ever saw one crazy 4 pawn sacrifice line that beat Stockfish (at lower depth) once.
In that game where Magnus Carlsen got beaten by Hans Niemann, it was a positional game that Carlsen played.
To date there has not been a Beating the Sicilian 4 published. Can the engine feedback be trusted here? Is the theory presented in Beating the Sicilian 3 still good or are parts of it out of date? If Beating the Sicilian 3 is no longer the best resource for White against the Sicilian, then what is?
1995 is an awfullly long time ago in in opening theory.
Some more recent books you might like. They're not the very latest word, but quite a bit closer to it:
Negi's three-volume 1. e4 vs the Sicilian set for Quality Chess
Sethuraman's repertoire for Chessable (I think the Sicilian work is in Vol. 2)
De La Villa and Illingsworth's "Dismantling the Sicilian."
I have been studying this great game, which ends in draw by perpetual after some ingenious sacrifices and a long line of only saving moves.
I have the depth 30 analysis on this game and it reports
https://www.chess.com/openings/Sicilian-Defense-Open-Scheveningen-Keres-Attack...7.g5-Nd7-8.Be3-Be7
White 98.2% accuracy, 1 brilliant move 23...Nxb5!!, 5 great moves, 1 mistake 6...Be7? best is 6...h6
Black 98.4% accuracy, 1 brilliant move 21...Bb7!!, 6 great moves, 1 mistake 13.e5 (GM Gallagher's recommendation) best is 13.a3
Unusually it reports that the only mistakes made in this game were made in the opening, but the game followed the leading theory at the time it was played in 1998/99.
On move 13 White follows GM Gallagher's advice in the book Beating the Sicilian 3, which was published in 1995 which has many good reviews and endorsements.
To date there has not been a Beating the Sicilian 4 published. Can the engine feedback be trusted here? Is the theory presented in Beating the Sicilian 3 still good or are parts of it out of date? If Beating the Sicilian 3 is no longer the best resource for White against the Sicilian, then what is?