The compute workloads for AI/ML are vastly different than mining, which means that the hardware requirements are totally different (and actually a huge capex in

The compute workloads for AI/ML are vastly different than mining, which means that the hardware requirements are totally different (and actually a huge capex investment). Individuals renting out their home compute for things like this suffer from the problem of supporting so many different types of hardware profiles. BOINC is a prime example of this, but it doesn't have a funding mechanism.