Ideally, it's close-- i.e. there is no ugly duckling. And where it's too close to call (i.e. you don't have a credible argument for why one person will have a greater contribution), I think erring on the side of almost-parity is the way to go, because stuff changes too fast and anticipating contribution is HARD. Today's shit-hot idea may be DOA in a month. Today's CEO may shift to VP Bizdev next year. Fundraising connections may come up dry when they are tapped. If any/all of that happens, then you have one guy saying, "Wait, he gets 65% of the company because of all of these contributions that really aren't working out. How non-awesome is that? And my good college buddy just asked me to jump into this NEW idea with 50% ownership... Hrm.".
Those are non-trivial risks for nearly-trivial (risk adjusted) dollar figures, IMO.
I see people disagree with IronYuppie (I do too - I don't think the CEO job is harder than the CTO job) but I'm not sure why it's getting downvoted.