You might be referring to Thrift. In that case, it's only useful if you want to do cross-language services. I don't think an early-stage startup needs this sort of architecture but I guess designing for scale never hurts.
1) Should you worry that you're relying on the infrastructure of another company? Will VCs worry about your long term prospects?
2) Would it be ok to use the developer toolkit until your business is a success, and then switch to your own code?
What are you trying to write, and how well does it overlap with the toolkit's intended purpose? Put another way, if you wrote your own, how different would it be?
Hopefully you're not imagining that FB for some reason decided to give away the tools that would be needed to create another FB. That's not what they mean by "toolkit".
Also, they're coming out with a bunch of new stuff towards the middle of may.
Clearly, it's hard to ask for advice on an issue that's not fully fleshed out.
It's just that so many webapps today need to have user authentication/profile management, it seems a shame to have to build your own for two reasons: 1) You're wasting your time building something. 2) People trust you with their personal information, I think people are more willing to trust, say, a facebook, then they are joestartup.com. I say this based on my intuition.
Just throw it out there, even if it sounds stupid in your own mind.