Equating that example to this poll is heading to very shaky territory. I can tgink of plenty of factors that would indicate both why that was an incredibly innacurrate poll (and polling method) and why a poll here on HN is probably going to produce at least soem form of workable data :)
Good on you to say "most respondents" rather than "most users." This voluntary response poll will tell us nothing reliable about most users of HN, even if more than half of all users respond (which is not particularly likely). See
http://www.edcallahan.com/web110/lectures/litdigest.htm
(you are probably familiar with the example, but this is posted for onlookers) for an example of a poll with a huge sample size that still got the wrong answer.