> This relies on the theory that startups goal is to grow a 100 times or die — so small continuous extractions by the executives are ruled out.

I interpret it as following: if the regular company earns $1 million/year, then an executive stealing $500k/year by making deals with friends will make investor's return much smaller. But in a startup, it does not matter -- if the it fails, it would simply fail slightly earlier. And if it succeeds, then it would earn $100 million/year, and $500k/year that executive steals will be a small, insignificant change.

(I am not personally convinced this is true, but this is how I interpreted that article)