> I’ve seen the idea that every minute spent on infrastructure is a minute less spent shipping features

This is why working in infra is so awful, people follow this principle for years and then invest a million a year in an infra team whose hands they tie.

Resilience is a feature, though I agree at early stage folks often need less than they think, a single auto-scaling group and load balancer in AWS or something similar isn't much heavier than a single linode VPS, except that it has substantially improved resilience.