Earlier quoted context omitted.
Can you do a back of the envelope calculation to figure out how large those "more valuable but also more difficult to monetize" effects would be? As I noted in the blog post you cited, most of this discussion is just meaningless verbiage. No one even attempts to quantify the the effects they claim are so important. (In the HN comments on that post, many people suggested such effects; the minute you crunch numbers the…
Non-empirically-validated numbers obscure a lot more than they clarify. The scarmig comment you link to shows that almost all of the effects you included were irrelevant and this whole exercise was worse than useless. Using this much precision at this level of accuracy can only ever mislead.
https://en.wikipedia.org/wiki/Fermi_problem
It's also about knowing what facts need to be true for a BI to work. Then we know what facts we need to verify to determine if it's true, in contrast to all the meaningless verbiage in this thread.
Also, Scarmig's comment was quite valuable. It showed exactly what the driving effect is - that's precisely why you should do a back of the envelope calculation. And as all the non-quantitative comments show, it's also why any comments suggesting an effect but not doing such a calculation are worse than useless.