I don't think we should trust the comp

change "I don't think we should trust the comp" to "I think we need to learn how to use the comp"

especially in the opening, stockfish is taking a lot of shortcuts and the evaluation has to be taken with a grain of salt, but...

even if the evaluation is accurate, stockfish is evaluating it against an opponent that plays perfect, another engine. there are positions that are really difficult for humans to play that stockfish has no trouble with and might even be winning for stockfish. So if you find a line that would put a human player in difficulty, does it matter that it's losing to a stockfish opponent?

Hmm......maybe even the comp makes SOME mistakes. I viewed a recent thread where it showed 1 e4 c6 as an inaccuracy.
I think it goes very well with that quote "If everything was perfect you would never learn and you would never grow".

The aim of chess is to checkmate. An engine's aim is to gain what?
The aim of a pawn storm is not explained by an engine.

youtu.be/RZYYUbhzTP8

Try to play it vs Stockfish and it will show you why it's -0.6

Well, hopefully this sort of thing will teach you how foolish it is to consult comps in the early opening. ;)

Always be skeptical, but not too much

IMO, any engine eval of a gambit that is beneath a full pawn deficit should be interpreted as just noise. When it drifts above that threshold, then maybe pay some attention.