They can't both be optimal. The probability that both are exactly equally good is so remote as to be ignorable.
Successful does not imply "no flaws." It's interesting to see this argument brought up in the context of biology. It's frequently applied to software, for example it shows up in basically every discussion of PHP.
They could both be optimal for their respective uses. Humans don't spend a lot of time on the seafloor, and cephalopods don't spend a lot of time in the open air. Certainly there are other tradeoffs to be made as well.
Successful does not imply "no flaws." It's interesting to see this argument brought up in the context of biology. It's frequently applied to software, for example it shows up in basically every discussion of PHP.