In terms of what your missing out:
You'll obviously lose DirextX9 hardware support.
You'll lose ATi's rather efficient anisotropic filtering implementation, so you can expect the GF4 to drop considerably in performance once anisotrophy is applied.
FSAA implementations are comparable at this point in time IMHO, though I'd give the slight performance and quality edge to ATi.
As with any ATi based graphics card it will most definitely provide you with a considerably more advanced TV-Out implementation on the driver level, and similarly DVD playback capabilities are quite a bit better.
You'll likely find the R9500Pro has slightly better 2D visual quality, though the impact of this varies depending upon whom you ask.
Future applications will most definitely strongly favor the R9500.... vastly more powerful pixel and vertex shading capabilities, along with full DX9 support.
Perrformance seems to vary depending upon which website you believe... some position is slightly above the GF4 Ti4600 whereas others seem to position it similarly to the Ti4200. Personally I'm inclined to believe it falls roughly between the Ti4400 and the Ti4600 in performance on most of todays applications.
Whether the above is sufficient reasoning to wait... well that's up to you.
Personally I'd probably hold off on the Ti4600, as the 9500Pro is simply a much more well rounded board with a clearly superior feature set and only narrowly falls behind in todays games... and should lead in future games as vertex shading, HOS (Higher Order Surfaces), and other other effects become more pronounced.
In the short term at least... for the next 5 months or so, the Ti4600 will probably give a narrow advantage in gaming IMHO, though if anisotrophy is applied I expect that to disappear rapidly.
So long as the R9500Pro starts appearing in the newt 2 weeks or so as it seems I'd lean towards waiting for it.