in order to push the innovation envelope, people must be working at, or at least have a deep knowledge of the current cutting edge of innovation. I agree, and

in order to push the innovation envelope, people must be working at, or at least have a deep knowledge of the current cutting edge of innovation. I agree, and that's a full-time job that has to be self-directed. Excellence is a nonlinear phenomenon. A person who excels in 10 hours of his time and subordinates in 30 isn't 25% as someone who gets to be full-time excellent, but possibly 1%. I spent 6 months at Google, a company that has some awe-inspiring R&D "centers of excellence" but is essentially just a bank if you don't get to work in one. Now, I'm sick of ragging on Google because I think I've overdone that schtick. On the other hand, with the incredible wealth of talent it has got, it can do a lot better. It's the best a closed-allocation company can be, and it has amazing people. But my experience there is instructive on the mediocrity that closed allocation command-and-control creates. If you don't have worker autonomy, you won't get creativity or excellence or innovation. Google remains awesome because the ~25% who get to be Real Googlers (they can, like Yegge, quit their projects and join better ones, because teams compete for them) have enough autonomy to do really good work and keep pulling the company, but if you're not a Real Googler, you're going to be working on 4th-Quadrant crap. (For the definition of "4th-Quadrant", go here: http://michaelochurch.wordpress.com/2013/01/01/fourth-quadrant-work/ http://michaelochurch.wordpress.com/2013/01/01/fourth-quadra... .) After that, I worked in a reputable (but diseased) NYC startup-- a VC darling-- with serious ethical problems at managerial levels (multiple people hired into the same leadership role; dishonesty about product direction used to keep people motivated). To my chagrin, the culture there was traditional MBA-style corporate bullshit: closed allocation, supernumerary people with fancy titles who didn't do much, extremely vicious politics (I resigned when asked to sign affadavits disparaging engineering old hands with too much equity). The Valves and Githubs excepted, these "hip" tech startups are not culturally innovative, and it's the prevalence of the old-style corporate culture that makes them all devolve into grey-goo mediocrity. Subordination is a genocidal killer of innovation, and the cultures of these "hip" tech startups are often not much different from large corporations in culture, creativity, innovation, or ethics. For every Valve, there are 100 companies that claim to be "startups" but are just as hierarchical and command-oriented. The hierarchy isn't as deep with only 100 people, but it's every bit as dysfunctional. If we want innovation out of these startups, we have to focus on culture. We have to stop funding culturally retarded people with MBAs just because they're "credible" and, instead, look for people who can lead very smart people (an extremely rare skill).