As long as it doesn't take initial development time away from the non-rift version, then it wouldn't hurt. As it stands, the dev kit isn't a good solution for gaming, and if the owners number in thousands rather than hundreds of thousands, then it just isn't worth it. It's a phenomenal and exciting experience, but it's a long way from being in the market position that Oculus would have you think it is, and I say that with a good deal of respect for them, I think it'll get there but it's not there yet.
Developing a good integration for the rift is a serious challenge. For some games it's simpler than others, but it doesn't get more complicated than a game that has visible first person body models (like the arm for the lantern) and object manipulation like in Amnesia games. A bad integration is ruinous to the sense of presence that a game might be trying to achieve, and given FG's focus on creating as strong a sense of presence as possible, it's safe to say that they aren't going to release a game with rift support that hasn't had a lot of development resource dedicated to working out a good UI/control scheme designed specifically for the rift.
I wouldn't lament a lack of rift support for AMFP, I trust FG & chineseroom to make the right call on it, whichever way they go.