These sort of computations generally just get fed bigger inputs as compute gets better. Also, plenty of threadrippers exist out there already, if you get acces

These sort of computations generally just get fed bigger inputs as compute gets better. Also, plenty of threadrippers exist out there already, if you get access to some cluster, it might have whatever type of chip in it. If I have access to a cluster with many 7995’s, I don’t really care too much about what’s available on the consumer side.