I'll add to the deadline section "the ability to actually ship." Beware of the brilliant computer scientist who is working on their baby, and sees the performance of their software as reflection of their character.
The software will never be "good enough" to release, and "no one will buy it" as it is. They'll run out of money before they actually launch anything.
I don't think I can stress this one enough. Not only will a team run out of money with a perfectionist co-founder, but the other obvious result will be that competition may ship a product before you do in what was previously a vapid market.
Should you happen to bring a perfectionist computer scientist on board, implement first (even if it's not DRY) and cleanup/refactor later. Always stress this message. Particularly Ruby programmers. You can remind them that once you ship, they always have the opportunity to clean it up during maintenance nights. You can make it work with a brilliant computer scientist, but always make sure he's focused on shipping rather than perfecting.
The software will never be "good enough" to release, and "no one will buy it" as it is. They'll run out of money before they actually launch anything.