Just to pick brains. I couldn’t help but map this discussion to statistical inference: Type I error and Type 2 errors (http://en.wikipedia.org/wiki/Type_I_and_t

Just to pick brains. I couldn’t help but map this discussion to statistical inference: Type I error and Type 2 errors (http://en.wikipedia.org/wiki/Type_I_and_type_II_errors#Type_I_and_type_II_errors http://en.wikipedia.org/wiki/Type_I_and_type_II_errors#Type_...) #1 Prediction: success Result: success true positive [excellent, no one believed huh] #2 Prediction: success Result: failure false negative [speculated reality? pun-intended] #3 Prediction: failure Result: success false positive [A splendid success but AI failed, so what.. err.. is there a point?] #4 Prediction: failure Result: failure true negative [It failed, but hey AI worked then, hmm.. whats the point] There is no good position except the rare event of a true positive. Banking on rare events isn't bad but, thats the fun of startup, isn't it?