But one is always guessing. Or always should be. You can't wait till you have complete information. You have to act on hunches.
Case in point: most of what Robby wrote was correct, even though it's based on a small amount of data. He's a smart guy, so he was able to extrapolate.
I think you have it exactly wrong. The mark of a great mind is not how few mistakes it makes, but the magnitude of the new things it discovers. Newton probably spent more time on pointless theological controversies than on physics. But we still consider him smart.
I have found that most disagreements are artificial; these are the result of a poorly framed discussion. I bet we agree, but are talking about two diffrent things.
I chose my words carefully. I did not say "the" mark of a "great" mind, but "a" mark of an "important" mind. I think we can all agree that Newton was a great mind, but I think the questions of what makes a great mind and an important mind are different.
I think there are many important minds in the myriad fields of human enterprise. These are people, like yourself Mr. Graham, who are able to understand their corner of the world and shape it.
I have observed that a lot of the people I see who are leading their fields have the quality I mentioned above. (These people have other qualities as well, but that exceeds the scope of the conversation.) They are in a position to lead, in part, because they have made a great many good decisions. As far as I can tell, the people who make good decisions are those who are not wrapped up in superstition and who are disciplined enough to make rational decisions whenever practical. Rigorously rational decisions require boundary conditions. It is important to know what degree of certainty is important. It is important to make a separate equally rational decision about how much time to spend thinking about the first decision (but be careful with this recursion.)
I can identity this pattern of decision making in your writing. You say that start-ups shouldn't spend a great deal of time worrying about their idea: "[if it seems good enough, then build it and find out.]" But, you rationalized that idea at length with the idea that implementing a project is less costly than researching the market. So, in this case, you are carefully making a rational decision to recklessly make other decisions. I think this suits my model (which is, of course, not certain).
You're both right. You need the balls to trust your hunches, to put yourself out there, even when there's limited data. But you also need the balls to throw your previous assumptions out the window, in the face of new data.
In the case of startups, it's true that everyone's advice is extrapolated from one or two data points. The key is to look at a lot of those data points -- read Founders at Work or listen to VentureVoice.com, for instance -- and try to find recurring themes and patterns.
This is part of the idea behind YC, right? I get that impression from your essays. These days there's less "At Viaweb, we..." and more "Out of the N startups we've funded, X have..."
