NVIDIA used CUDA to establish industry-wide vendor lock-in on GPU compute. Google uses TPUs to try and establish an industry-wide vendor lock-in on deep learni

NVIDIA used CUDA to establish industry-wide vendor lock-in on GPU compute. Google uses TPUs to try and establish an industry-wide vendor lock-in on deep learning. Same old, same old.