One way this plan could fall through is if LLM tech progresses to the extent that free and open source copilots, run locally, can give result that are just as good.

MS might be in trouble then.

Performance is not great but apparently it’s not optimized at all as of right now.