[082349872349872]

But the "mostly waste of time" is true: you don't need all the advice, you only need the 10% (1%?) of it that's actually any good.

(The biggest hurdle is getting enough expertise that you can ignore the dross?)

The flip side of this is you can't just read advice; you have to try it (and maybe even try its opposite). Expertise is not a spectator sport.

[sasha_fishter]

I agree, but do you think pursuing that 10/1% is worth spending time reading books/articles?

[082349872349872]

Yes. For instance, I've read the first and last books mentioned in https://www.allthingsdistributed.com/2020/07/the-os-classics... and wholeheartedly recommend them (esp. Gray&Reuter!) to anyone with a systems bent. (I've also read several others mentioned there and based on their quality I'd be willing to pick up the ones I haven't read)

I guess it's a bit of a chicken and egg: you need to skim ruthlessly, ignoring (à la Sturgeon's Law) most of what is proffered, skim what you don't ignore (index, table of contents, and a few random pages within), but when you do find "the good stuff", stop, read it, and probably also read works it refers to. Unfortunately for the initial filter to work, you can't be a complete neophyte, but luckily practice helps a good deal there.

Looked at from the other way around ("reversing all the arrows"): whenever you find a work that is helpful for expertise, stop, soak it up, and explore its neighbours. This will be time intensive, especially the practicing part, so of necessity you will need to leave the vast majority of things unread...

> Ved å studere mesterne, og ikke deres elever — NHA ("read the masters, not the students")