Reading the article "The Interview with Y Combinator That’s Not" earlier I thought there was an interesting angle to be explored.
During the interview, the DirectedEdge team got a question, "So where’s the rocket science here?" This clearly attaches some value to the underlying implementation and algorithms.
How critical is having some super-advanced software tech to startup success?
One of the reasons I originally posted the question is that I'm particularly interested in one of the ideas that YC mentioned in the startup ideas that YC would like to fund, enterprise software 2.0 (http://ycombinator.com/ideas.html see number 5).
My experience in that space is that there are a number of companies that are generating huge revenues using software that could be improved via the "build a better mousetrap" methodology. But basically none of that space requires rocket science, unless you include the ability you have as a startup to use more powerful languages and tools than your standard corporate behemoth.
Now, that said, using more powerful tools might free you up to innovate in ways other companies can't, similar to how PG has touted Lisp as a big strength for Viaweb . . .
Secondly, the statement points to how you protect your IP. If it's something that's easy to make the first time, then it's even easier to duplicate. How do you protect your idea without any kind of "secret sauce"?
Wrap it in a business model that's somewhat robust to competition. Photo uploading is not technically complex, but the business model to support free is more of a barrier to entry.