Qualcomm open-sources Mojo compiler to avoid CUDA lock-in
Three weeks after closing its Modular acquisition, Qualcomm open-sourced the Mojo compiler on GitHub under Apache 2.0 with LLVM exceptions, ending years of a closed compiler. Mojo is designed to compile one AI codebase to run across Nvidia, AMD, Apple, Qualcomm and other accelerators without CUDA lock-in, though parts of the stack had been open earlier.