[m-p-3]

I'm wondering if something similar could be done but for ML, using a distributed network of GPU to train a ML model in exchange of a payment for using the resources.