1. subjective belief 2) experiment is commitment 3) engineering orientation vs hypothsis testing (academic vs ); randomized trial ex post = prior (public demonstration; won't do it) how is experimenting different from (deliberate attempt to gather information) science experiment VS A/B testing VS Raffel rait (whose primary goal is succeed; because we'll be learn as we) deliberate activity to generate information (busines plan, ip strategy, potential customer) high fiedelity view point on private return - bias? (stage finance model; no cap ) everyone new thing; run after (john horton; equiliibrium = entry high as no expectation of competition) end up diverting money (camer jpe enter or not) experiments you're gonna run (money and) are assumed; discursive (information issues) eric ries (replaced "listen to investor" -> "should do experiment" ; created ability to ) eric: different belief (affiliate with ppl you agree with) - body 🙋‍♀️1: compared to software where experiment cost dropped, this phenomena wasn't shown in deep tech more money is available, investors would fund risky ideas, early stages are anticipating (later stage would ses cheaper) highest possible s1 (true positive), lowest possible s2(true negative) run my business (C for the ) lambda0, lambda1, principle (we're designing unbiased experiment; expected of my posterior (integrated of) = prior) mu/1-mu = 1-lambda0/1-lambda1 you don't have a choice (your prior determines a mix) mubar = C/V biased experiment to be run by entrepreneurs expected posterior = prior (learning cannot be biase; if you poseri; if your expection is differet from ) informative experiment doesn't depend on prior lowbar는 fn최소화; subjective willingess