ParentFull threadZambyte·Who can afford to service game streaming when those GPUs could be used for LLM inference / training?View on HN