Idk, it's fuzzy, obviously, and a moving goalpost, but in the end it comes down to this: do I buy/rent a bigger server? Or do I put more engineers to the task o

Idk, it's fuzzy, obviously, and a moving goalpost, but in the end it comes down to this: do I buy/rent a bigger server? Or do I put more engineers to the task of replacing the naive algorithm. The latter is Big Data and usually the case when you run out of RAM, not out of disk. And the first one is an underrated option.