Yeah, suggesting the prioritization of focus over relationships is like suggesting the prioritization of focus over sleep or exercise. Or even of one area of focus (say, math) over another area of focus (say, knowledge of a specific codebase). You can do it, maybe situationally it will be the
right thing to do... but any specific "greedy" focus choice is going to have some substantial costs.
If you're going to approach focus like an engineering problem, you probably want a minimax strategy: figure out what keeps you functioning/healthy and devote minimum adequate focus there, figure out the reward space you want to explore and devote near-maximum resources there. And leave a little bit of slack baked in for when attention is overscarce or for when you need perturbing input. Digital distractions can be that slack. Or they can be your dump stat (he says, while posting on HN :/ ... ).
Most people will find that a handful of close relationship plus some web of weak ties fit in both the "keeps you functioning" and "reward space" categories. Which both underscores the importance of relationships, and the challenge: you are probably going to have to balance frequent not-especially-rewarding maintenance work, and regular (but longer period) reward-space exploration, and feeding slack time to it.