A large part of your response seems like you think you're arguing with me,
Were I arguing with you, I'd tell you why you were wrong, not why you were right.
I don't see why not, as long as your definition of "people who aren't right" isn't about race, gender, or sexual orientation.
Here's where you are wrong. It is easy for any significant filter on subjective qualities to be interpreted as illegal.
That is, if your objection is something like, "This won't work once you have black employees!" then you should fuck right off.
First, that very well could be a large part of the reason your 40 person company doesn't need a CEO (a monoculture, either way), but you would certainly be a fool to say it out loud. The problem comes when your "people who are right" don't happen to include any black people for whatever reason. The assumption is you intended what you just said, and we see your own reaction to that.
But labor laws don't really prevent other forms of discrimination, based on things like incompetence, lack of qualifications, or bad behavior.
But those are not the things that you'd need to select for to make your 40 person CEO-less company continue to work at 100 persons. It's not just an issue of qualifications or "bad behaviour." It's "able to work well in a leaderless environment." Not just that, but "work well in THIS leaderless environment". That means you may wind up selecting someone from a majority who is less qualified over a minority with much better qualifications, or vice versa. Or selecting only men, or only women. THAT is a recipe for an EOE lawsuit whether you're actually discriminating against a protected class or not.
Statistics are how this stuff is measured, because looking at each case individually by regulators is too hard. It applies in hiring, and Title IX, bank loans, and all kinds of places. Anything that makes the stats unbalanced is prima facia evidence of wrongdoing, even if the reason is as simple as "no women applied." Obviously you discriminated in the job announcement to dissuade women, then. If your selection criteria for "this leaderless workplace" happen to result in a statistical anomaly for any reason, you're a target.