[karmakaze]

Maybe the thing to realize is that the idea of BEST is very much in the realm of diminished returns. Being able to internalize that and the recognition that there is no single best, but varies for each new context will show that it's not a 'once and done' thing worth knowing. Don't make secondary goals a primary objective.

I find that jr devs are preoccupied with learning what's best for each category of thing. A more seasoned dev recognizes that everything's a trade-off, makes appropriate satisfactory picks and carries on pragmatically to get something end-to-end off the ground. Then they can iterate on the next most important area do improve.

The best way to get into that pragmatic loop is to have actual users providing feedback. It's much harder to justify optimizing something with little impact while there are real bugs or glaring missing features in the queue.