it's a serious question! i have never written a (successfully merged) spec before.
seems like the key goal is to disambiguate any possible misunderstanding of the important implementation details while leaving the rest of the decisions up to the implementor. does this require two different individuals to implement it? if an llm is implementing it, are two different prompts sufficient?
seems like the key goal is to disambiguate any possible misunderstanding of the important implementation details while leaving the rest of the decisions up to the implementor. does this require two different individuals to implement it? if an llm is implementing it, are two different prompts sufficient?