Sure. Just realize that PC != console.

In terms of proving your ideas, go hog-wild with whatever you can achieve on PC.

In terms of making a potential port to consoles easier on yourself, you should try to become more aware of the features, limitations, and expectations of console games. This can range from "I know that next-gen consoles have 3.5GB+ of usable memory" to "I know that next-gen console GPUs are roughly equivalent to X graphics card" to "but I also know that next-gen console have lots of relatively slow CPU cores" -- it also means having some idea of what extra hardware the consoles will have to assist you in certain tasks, and that certain things that work on a PC monitor at 24 inches--like small text--doesn't work on a large-screen TV at 6 feet, even in 1080p.

If current rumors are to be believed (and it seems to be that we'll find out about the PS4 on the 20th of this month), both next-gen systems will be sporting 8 AMD jaguar cpu cores @1.6Ghz, 3.5-5GB of usable memory, and DirectX 11.1-level AMD graphics with performance between 1.2 and 1.8 tflops. If you were to build a PC with the purpose of approximating these characteristics as much as possible, you might shoot for something like an AMD FX 8530 (disabling L3 cache and under-clocking to ~2Ghz), 8 GB of fast DDR3, and something like an AMD 7770 or 7830 GPU. You'll never replicate a console with a PC for a multitude of reasons, but the system described would be a not-unreasonable approximation. Of course, if you aren't aiming to push the boundaries, you don't need to concern yourself with such things.

I just think it would be a sunken cost to buy a ps3/X360 dev kit as they are nearing the end of their life cycles.

Well, its not like you can just buy a console dev kit anyways, even if you do have a demo and even if you're a recognized studio -- Many real AA-level studios own precious few, if any, dev kits on their own. Many times kits are leased or lent to a studio by their publisher, who actually owns the kits. Even then, to have a dev-kit on your premises generally requires that you have a real office, and can prove that your physical and IT security is sufficiently guarded.

In general, unless you have actual industry contacts (people involved with the publishing side, not some game tester you went to high school with), I wouldn't really recommend setting out with the primary end goal to be a publisher pitch. A publisher won't even open an unsolicited pitch or demo just to avoid any potential legal claims made by outside designer--bottom line, unless you've been given the name of an actual person to submit to, through official channels, then you've got no chance that they'll look at it.

If I wanted to parlay my idea into a publisher deal, what I'd do is set out with the goal of producing 10-15 minutes of good, representative content and gameplay, with the purpose of crowd-funding development of a full, if-short, game -- maybe 4-6 hours of playtime. It's *that* product which I would use as the basis for my publisher pitch, and at which point you might be able to roll into a publishing deal for the "real" version of that short game, or even into a different IP entirely. It's a long, long road to getting in with a publisher -- if you think it's gonna happen in one step you best prepare for disappointment. But by taking the right steps and being persistent, its still doable, even as the AAA industry becomes more and more insular.