I found Tim Oren's piece on the emergence of a two-tier software venture useful.
He argues that software startups will get going on the cheap. They'll develop their first product with with open source tools and will send nonessential work to India for cheap/rapid development. They'll use founder/angel money to build a useful and differentiating application. Then market it through word-of-mouth, Google Adwords, and other cheap options.
If things pan out from there, VCs can step in to provide funding to build a sustainable business.
This is more or less the approach we're taking with TripInvite. We have a good concept and a good team. Before I visit any VCs though, I want the software to be launched, well tested, tweaked, and to have our first 100 customers. Since we're developing on the cheap anyways, there's no reason not to do this... remove as much risk as possible beforehand. If we were developing nano-something-or-others, this obviously wouldn't be the case. The technology there is still super expensive, and bringing VCs in at an early stage still makes sense.
Tim argues that the commoditization of IT is what drives this model. Its not the technology, or access to technology, that makes you special in software. Its about having a good idea, doing something that's hard, and executing well.
Anybody with a good idea can start a software company. We 100% outsourced our beta website, with a significant amount of code behind it, to Bangalore. Neither my partner Dave nor I knew how to write 1 line of code.