I get you. But here's the catch, and that's also valid for HN: Is crowd validation always the best measurement of the quality of an article? I agree there are a lot of running in circle conversations on HN. Like "Apache is slow, use Node.js", or its contrary "Node.js is bad, use {insert something else}". On HN there's a great deal of know-it-all attitude and most of the posts could be summarized by "Your choices of X is stupid, I chose Y and it's better, therefore, ' Why X is bad and you should use Y' blog posts ".

Anyway, 37signals could be the Kim Kardashian of Software Engineering, very popular and very successful, lots of people following the Gossip. Popularity does not imply quality, that's why civil disagree-ing is useful, we need safe guardians of the quality, even if they seem obnoxious when they criticize what we like.

PS: 37signals is NOT, in my opinion, the Kim Kardashian of Software Engineering.

The only people who know whether they are mere celebrities are those who have seen their code.