Yeah I guess that might work for some workflows, we try and keep local as close to prod as possible, and how we build local artifacts that we test as close to how we build prod artifacts. Giving us a fighting chance of being able to repro prod bugs for example, if a team was making a different set of trade offs that would work.

Thinking about it more, we have had the problem of docker desktop using a bunch of memory and slowing productivity, so shipping that giant vm elsewhere would solve for that, can I spin up 16gb+ namespace/env in okteto?