Comment by AndrewKemendo

I’m leading engineering at a new start up and I’m doing absolutely everything I can to avoid touching the cloud at pretty much any cost.

We’re going to be running all of our own dev/test GPU servers with only inference hosted outside our cluster.