Isn't it a pretty big caveat to exclude people who've "worked for a startup and it's failed", given how large a proportion of startups fail? I mean, sure, if you only consider employees of startups that ended up with large exits, things look a lot rosier than if you look at the whole set of startup employees.