'Studying with Modern Engines'

The September 2022 Chess Life had a topical article, 'But Does It Actually Work?; Sadler and Doknjas on improving by studying with modern engines' by IM John Watson. The master book author and reviewer discussed two books:-

• Sadler, Matthew. The Silicon Road to Chess Improvement. New in Chess, 2022.
• Doknjas, Joshua. The AI Revolution in Chess. Everyman Chess, 2002. [sic; 2022?]

The article is available online at Watson Book Review: Sadler, Doknjas, and Modern Engines (uschess.org). Watson explains,

To grossly oversimplify, the older-style engines (like Stockfish Classical, with what he calls 'hand-crafted evaluations') have continuously improved and are generally superior in calculating outrageously deep and ingenious tactics. Nevertheless, the neural network engines like Leela Zero with self-learnt evaluations can play a more profound and effective strategic ('positional') game, which tends to outperform the calculating monsters.

What can be learned from these engines, especially the NNUE engines? Watson again:-

These books convince me that engine study can lead to improvement, but generally in fairly narrow and specific contexts. First, by finding exact orders, well-timed maneuvers, and successful plans in the opening, as is practiced by every leading player in the world. More generally, in discovering typical maneuvers in certain structures and better evaluation of contrasting strategies -- for example, certain pawn sacrifices or flank attacks.

What does that say for chess960? The opening is different for every game and the phrase 'typical maneuvers' is meaningless. Perhaps it's better to ignore the engines completely.