Smart people don't play monkey games. If you want a monkey at the very least be honest about it. If you want a rockstar you better have a damn good reason for them to come to your company rather than just start their own. Playing monkey games to see if your possible new team rockstar is good at reading minds through osmosis is the first reason they won't even look at you. Better stick to brogrammers, however I doubt you even see many of them.
As for your secondary point about the quality of our team, I encourage you to read our engineering blog (http://www.thumbtack.com/engineering/) and look at some of our open source contributions (https://github.com/thumbtack). I think you will find more than enough material there to see what our engineering team is all about.
One thing I do agree with in the original post is that the resume is a really bad way for me to learn about candidates, and most candidates resumes are much too long. A blog, open source contributions or stackoverflow answers are much more helpful and really make candidate stand out from the crowd for me.
Within the last couple months we also started requiring all candidates to do a coding challenge before applying (http://thumbtack.com/challenges), and it has been an epiphany for us. My current process is to look at the challenge submission without seeing anything else about the candidate, grade it, at which point I'll then look at the resume to see how many years of experience (I have higher expectations for the solution the more years of experience the candidate has). For borderline submissions I'll dig deeper into the resume, but 90% of the decision comes from the challenge submission and maybe 10% from everything else.