RISC-V is it about to be transformed by AI tools?
IT seems to me that good part of underlying fundamental assumptions for RISC-V just got rejigged if not outright smashed by latest AI advances.
They seem to be on verge of being capable of reimplementing whole compiler/assembler/debugger backends to fit the ISA.
If so, how many of RISC-V fundamentals are still needed ? Maybe the time is coming for RISC-VI or something like that to be optimized for new reality ?