I just think let's not do it hastily otherwise we may port to a less secure system. These new algorithms are proven mathematically (possibly), but actually designing classical software to implement then perfectly may be so full of bugs and attack vectors. Let's be careful as this is all new and is all from a point of fear, not strength.
Now imo, if we ever have to do it, I'd say let us do it after quantum computers have proven their threat. Since you agree with the energy signature argument, then you see that if a QC that can actually break a Bitcoin wallet is designed it will both be major news to that computer company and a way to rally Bitcoiners to move coins. The initial loss will only galvanize Bitcoiners better, while it will rally a lot of people to sympathatize with us against the big bad monster.
Also, as I have posted on Scott Aaronson's AMA page, quantum computers will be useful to Bitcoin so I'd say we call their bluff. Motivate the QC scientists to actually build them.
Otherwise, nah. It's way out there. Just like nuclear fusion.