It's as if there are two ways of valuing companies, two competing schools, and their methodologies are opposed, arguably in outright conflict -- they can't both be true. One is the valuation on so-called fundamentals, the other is the bet that this will one day become one of those extreme outliers whose ROI is such a high multiple that it makes up for the losses on other failed bets. They're both valid models, like Newton or Einstein, that produce correct (enough) results, but in different situations and at different scales. The trick is to know when to use one model to make your decision, or the other. To carry this forward, 37Signals/JF/DHH clearly prefer/bias to the Newton method, in this analogy, and YC & SV VC the Einstein. Or to change the analogy, the former models are based on arithmetic, the latter on statistics. They both can work well but each in a different context with different endgame goals.

(And thanks for the reply. I'm a fan of your work.)